Mellon Foundation and Carolina Population Center

"Geographic Accessibility and Site Suitability: Factors Influencing Settlement Patterns and their Sustainability in the Ecuadorian Amazon"

Stephen J. Walsh
Carolina Population Center
Department of Geography
University of North Carolina-Chapel Hill
1998-1999

(1) INTRODUCTION

In the currently funded NASA project (1998-2001), "Agricultural Colonization in the Ecuadorian Amazon: Population, Biophysical, and Geographical Factors Affecting Landuse/Landcover Change and Landscape Structure," R.E. Bilsborrow and S.J. Walsh, Co-PIs, examine the human and biophysical dimensions of landuse and landcover change (LULCC) associated with spontaneous agricultural colonization in northeastern Ecuador. A satellite time-series of landuse/landcover (LULC) change and GIS thematic coverages of biophysical gradients and spatial patterns are linked to a scientific population sample of georeferenced farmhouse data on LULC and socio-economic characteristics at the household-level for 1990 and 1999. Image processing for LULC characterization and spatial analyses of landscape structure will be used to assess the rate and nature of LULCC and to model the effects of LULC, secondary plant succession, and land fragmentation on carbon budgets and assimilation rates for specific landscape strata and study area locations. Statistical models will be used to estimate the demographic, socio-economic, biophysical, and geographical determinants of farm and community-level LULCC. The degree of agricultural extensificaton and intensification and rates of deforestation will be documented at the household or finca level, the sector or community level, and the regional level. The research combines social science survey methods with environmental modeling, landscape ecology, and human ecology in a GIS to increase the understanding of LULC dynamics and the forces influencing deforestation.

The overall goals of this newly proposed research, while rooted in the ongoing NASA work, are distinct in a number of important ways, and can be summarized as follows: (1) explore additional population-environment relationships in northeastern Ecuador by considering the influence of environmental site suitability and geographic accessibility on the sustainability of household farms and the development of adaptive strategies by household members to increase their absolute and relative advantage for economic viability in the face of a diminishing resource base and a remote and disconnected geographic location; (2) examine the differences in the spatial extent, size, and geographic pattern of land clearings and related landscape disturbances associated with spontaneous settlers and indigenous people by studying two selected pilot areas within northeastern Ecuador, Yasuni and Cuyabeno, and by characterizing the dominant landuse/landcover types, plant communities, and land clearings that surround the Yasuni National Park and the Cuyabeno Faunal Reserve. These national conservation areas overlap with oil exploration and agricultural settlement areas, as well as known indigenous tribal reserves; (3) train members of EcoCiencia, an Ecuadorian organization participating in the NASA research, in the collection, processing, and interpretation of spatially-explicit environmental and population data and their integration through GIS approaches; and (4) develop a pilot database through additional data collection and processing to support a set of hypotheses, revised and articulated through this proposed research, to serve as the basis for subsequent proposals.

The GIS database, satellite time-series, and the population surveys developed to support the NASA-funded research will underpin this newly proposed initiative, but they will be augmented by additionally collected data and separate analyses to explore new areas of scientific inquiry that effectively leverage the developed informational infrastructure and knowledge-base into additional hypotheses framed and beta-tested through this pilot research to respond to forthcoming proposals that are being considered for NASA, NSF, and NIH.

(2) Purpose and Significance

The objectives of this research are to examine relationships in selected pilot areas in northeastern Ecuador, Yasuni and Cuyabeno, related to the broader area of concern, population and the environment, by assessing the sustainability of household farms or fincas as a consequence of (a) site suitability of household farms for agriculture and the related environmental degradation of lands through decreasing soil fertility as a consequence of cultivation and associated agroforestry, (b) variation in the nature of land clearings and in the definition of site suitability and sustainability as a consequence of landuse practices associated with indigenous peoples versus spontaneous settlers; and (c) geographic accessibility of farms or fincas to centers of supply and demand by taking into account geographic location, measures of impedance (e.g. Euclidean distance and route distance), and spatial interactions with other fincas and elements of the local and regional community (e.g. towns, clinics, churches, and schools). Each of these research objectives offer the opportunity to expand the existing NASA grant into collateral areas of study that remain rooted in the ongoing NASA research but depart from it in significant theoretical and methodological ways.

The hypotheses that shape the above research objectives, respectively, state that (a) environmental site conditions (e.g. soils and topography) vary within a three-dimensional matrix (i.e. x, y, and z dimensions) that produces a comparative advantage or disadvantage for subsistence and commercial cultivation. But, because of the declining soil fertility in tropical environments associated with sustained agriculture and exacerbated by poor initial site conditions and continuous cultivation, environmental degradation results over time thereby requiring adaptive strategies by households to maintain their economic viability. Such strategies may include crop rotation, land conversion (e.g. cycle of landuse initiated by deforestation of primary forest and conversion to cropland, then to pasture, and then back to forest through secondary plant succession), intensification, labor-sharing, and land abandonment, sale, or subdivision; (b) the spatial imprint of settlement on the landscape varies as a function of culture and ethnicity, subsistence or commercial cultivation, crop types grown, resource endowments, size of land-holding, and social and environmental policy among other factors. Areas bordering conservation areas will be examined relative to indigenous land-clearing patterns versus spontaneous settler patterns by emphasizing differences in landuse types, forest cover-types, and characteristics of land clearings. Indigenous people are more likely to produce land-clearings of a smaller size, alter the biodiversity of the forest to a lesser extent, and generate less of a geographic reach on the landscape resulting in a less consolidated and clustered distribution occupying a more limited areal extent; (c) geographic accessibility of farms or fincas has undergone a change since 1990 because of the expansion and improvement of the road network within the region. New roads have altered the calculus of geographic isolation or remoteness by altering the comparative advantage of sites through improved access. Fincas initially located away from roads were regarded as less valuable, fincas adjacent to roads were considered more valuable, and land parcels within fincas that were positioned adjacent to existing roads exhibited a comparative advantage and inherent locational worth and therefore were initially developed through deforestation and subsequent cultivation. Since 1990, new roads were developed, existing roads improved, and new points of economic opportunity, such as towns, were developed that changed the accessibility equilibrium and hence altered the pattern of deforestation and the nature of agricultural extensification and intensification at the household-level in these newly accessible sites. In addition, new sectors (i.e. a collection of household fincas) have been developed since 1990 that are now associated through geographic accessibility via the current road network. Fincas and sectors developed since 1990 are not part of the NASA-related social survey. Therefore, this study will also collect basic geographic and landscape information on "new" sectors and their corresponding fincas since 1990. Their topologic relationships to 1990 sectors and their individual household fincas, community features, and road network will also be assessed.

Satellite imagery, a GIS database, and on-site data collection, geographically referenced through global positioning system (GPS) technology, will be used to address the above issues. Satellite data have been successfully used to compare patterns of LULC influenced by local factors, population characteristics, and exogenous considerations. Among the factors that have been assessed include settlement history (Allen and Barnes, 1985), population size and distance to the nearest market-place (Behrens et al. 1994), socio-economic and cultural differences within the population (Moran and Brondizio, 1998), and landscape organization of LULC patterns (Walsh et al. 1998). The intensification of LULC also has been found to be associated with greater deforestation and a more heterogeneous landscape (Dale et al., 1993). They argue that deforestation is a socio-economic process that can be studied through social data collected on-site, spatially-explicit data on soils and vegetation, and models that link socio-economic and ecological processes. Explicit is the integration of spatial analytical tools to configure data for compatibility, assess landscape change through a remote sensing time-series, and perform spatial analyses of landscape form and function.

Often the missing part of the equation in this type of analysis is the linking of deforestation to the plot level. In the NASA research, individual households sampled in 1990 will be revisited in 1999. The sampled households and their associated lands are spatially referenced within a GIS. In this proposed research, landuse/landcover change associated with indigenous lands will be spatially referenced through satellite data, site conditions of spontaneous settlers are documented at geographically referenced fincas, roads are also place within the same geographic reference system, and the 1999 road updated and also spatially referenced. In short, the research, both the NASA work and this newly proposed research, operates within a spatial context that places landscape elements and social data within a co-registered framework sensitive to spatial and thematic inquiries operating at fine and coarse grains.

Political ecology, human, ecology, and landscape ecology offer a theoretical context to this research, respectively, by considering deforestation in the context of larger social, economic, and political relations (O’Brien, 1997); recognizing that important feedbacks exist between people and the environment (Bilsborrow and Geores, 1992); and associating landscape form or pattern to landscape function by integrating social, biophysical, and spatial domains whose effects are scale dependent (Walsh et al., in press).

(3) Study Area

As described by Bilsborrow and Walsh in their 1998 NASA proposal, the northeastern Ecuadorian Amazon basin is comprised of five provinces: Napo, Sucumbios, Pastaza, Morona Santiago, and Zamora Chinchipe. Within this region, the population is overwhelmingly rural and consists largely of migrant settlers that are characterized with high fertility and high mortality levels. In 1990, the census population for the region was reported at 371,000 of which nearly three-quarters were rural. The region also continues to experience a high rate of population growth, over 5% per year in the last intercensal period, 1982-90, double that of the country as a whole. Almost one-half of its population was born outside the region, two-thirds coming from the Sierra, and most since the mid-1970s. Small farm households are considered the primary driving force of conversion of primary tropical forest to other land uses. Government policies have encouraged this migration, as the Amazon region has been perceived as an area with almost infinite space and resources.

Access to the northern Amazon was initially made possible by the petroleum boom, which began in the early 1970s and led to road construction to support petroleum exploration and extraction. Access and land occupation in Napo and Sucumbios provinces have been in part by-products of petroleum production. Migrants follow the new roads into the region, which lead to the major towns in the study site -- Lago Agrio, Coca, and Shushufindi, occupying lands along the roads in a landuse pattern known as respaldos, layers of landholdings developed parallel to the main road. The first migrants tended to occupy the first layer, or linea (line) along the road; subsequent settlers occupied lines increasingly farther from the road. This process has continued up to 14 respaldos in the 1990 population survey area, though two to five is far more common.

Land quality and soil type vary widely across the sampled fincas: 46 percent of the settlers farm what they reported to be good soil; another 24 percent farm acidic soils of low fertility; 5 percent farm lowland alluvial soils with poor drainage; and 25 percent farm a combination of soils. Two-thirds of the farmers surveyed in 1990 reported declines in yields on their plots during their time in the Amazon. The sedimentary soils that predominant throughout the study area are of low natural fertility and experience a rapid decline of nutrients with use. In addition, the structure of the soil is easily compromised when subjected to disturbances associated with cultivation.

Due to the lack of a dry season and high annual precipitation (2,800 mm/year) in Ecuador, the slash-and-mulch system is a rational adaptation and differs from that of most other Amazon frontier regions. The polyculture system integrates annual crops such as corn and rice; semi-perennials such as plantains, bananas, and yucca; and tree cash crops such as coffee and cacao. Eighty-five percent of the households surveyed in 1990 marketed some crops, mostly coffee.

(4) Research Design

(A) Site Suitability

Previously collected GIS coverages of water and transportation networks, terrain characteristics, soil conditions, and the like will be linked to the satellite data to ascertain the site conditions associated with individual household farms. In short, each household will be assessed relative to land suitability characteristics for agriculture. Site suitability indices will be developed at the household and sector levels that integrate soil conditions (e.g. fertility levels), terrain characteristics (e.g. upland versus lowland topography), hydrographic features (e.g. surface water accessibility and the spatial variability in soil moisture conditions), and landuse/landcover types. Continuous surfaces of environmental factors, developed in the GIS to support the NASA-funded research, will be used along with satellite-derived measures of landuse and landcover for 1998.

Because of the nature of soil conditions, agricultural practices are subject to the cycling of crop types as a consequence of a relatively rapid declining soil fertility. Immediately following deforestation, soil fertility is relatively high, but after approximately 5-7 years the soil fertility is substantially reduced necessitating shifts in crop type to less demand types. Market conditions for crops, soil fertility reduction curves, and the cost and availability of agricultural inputs particularly fertilizer collectively influence decisions related to initial plantings following deforestation and the possible change in cropping patterns with time for economic sustainability. In most situations, crop rotation includes a fallow period or a return to forest through secondary plant succession. Satellite digital data will be manipulated through vegetation indices (e.g. the normalized difference vegetation index) to assess the relative differences in plant biomass across space and time associated with land clearings and deforestation; land classifications of cover types and their changes through time will also be derived. Absolute measures of plant biomass can be developed by linking spectral responses from the satellite to measured (or reported) or estimated yields at specific land parcels. The site suitability index will be derived through multiple regression techniques and validated through direct landscape observation and site measurements, referenced in space through GPS technology.

Satellite data sets will be used to examine the spatial imprint of land activities associated with indigenous peoples located adjacent to two national conservation reserves and compared to land patterns associated with spontaneous settlers within the same general locale. A satellite time-series will be used to examine spatial shifts in land clearings, the nature of the land clearings, and the sustainability of those clearings over time and space. On-site field data will document the types of land-uses practiced by both groups and their spatial and compositional characteristics. Ground control data will be collected to validate the satellite image interpretations. GPS technology will be used to spatially reference the land clearings, landuse types, and plant community types for accuracy assessment of the satellite image processing.

(C) Geographic Accessibility

Below briefly describes how accessibility will be examined within a spatially-explicit framework of a GIS. Road segments added to the network will be updated through remote sensing techniques, and road type changes will be assessed in the field and through remote sensing techniques. Community features (e.g. point locations of town centroids, churches, schools) will be captured in the field through GPS technology.