Roads : a proximate or underlying
cause of deforestation ?
Marie Scouvart, Eric Lambin et al.
Université Catholique de Louvain
Fonds National de la Recherche Scientifique
IIId LBA Conference, 27-29th July, Brasília
Special Session 7 - Roads, land use and land cover changes in Brazilian Amazon
Summary
• Introduction : general context
• Theoretical perspectives on the role of roads
• Research design
• First preliminary results : cases outline
• Methodology
• Policy implications
General context
Nearly 90% of deforestation in the Brazilian Amazon between
1991 and 1997 occurred within 100 km from the five major road
networks (Alves, 2002)
National demand for land Policy decisions to Migrations Conversion of forest
High agricultural suitability extend the road to agricultural land
of some forest areas network in those areas
Or
Expansion of local logging Need to link these active Construction of new roads
or agricultural activities in production areas to
some forest areas existing markets
Or
Construction of a road Spatial redistribution of New deforestation
for reasons unrelated to population and activities
land use in the forest
Attraction to a given Deforestation
Or location of a preexisting
demand for land
Research questions and hypothesis
• Objective
understand the role of road development on deforestation and
livelihoods under different natural and socio-economic conditions to
improve scenarios of future forest-cover change
• 3 questions
– are roads a proximate or underlying cause of deforestation ?
– do roads mostly affect rates or location of deforestation ?
– are roads endogenous or exogenous factors of deforestation ?
• Main hypothesis
in most cases, roads are only a spatial determinant of deforestation at
the proximate level. The quantity of deforestation is mostly determined
by macro-economic factors, national policies and/or patterns of migration
at the national scale.
Roads are catalysts of deforestation
Research design
• Comparative analysis of local case studies
– identify the factors that control the impact of roads on deforestation
– exploit in a comparative framework the knowledge acquired through local
scale case studies
• Case studies selection
– based on available literature and active research teams
– cover the full range of impacts of roads on deforestation
• Data collection
– published case studies
– formal expert consultation
• qualitative information, expert judgement
• quantitative data
– fieldwork
• key actors, institutions, etc.
ALTAMIRA - PA
URUARÁ - PA MARABÁ - PA
GUARANTÃ DO Adams, PARAGOMINAS - PA
E. Moran, E. Brondizio, R.T. T. Evans
NORTE - State University]
R. Walker, M. Caldas, S. Aldrich, E. Arima [Michigan MT
[Indiana Nédélec
V. University] / Cifor / IPAM] C. de Souza [Imazon]
P. Pacheco [Clark University[Université Rennes 2]
SÃO FÉLIX - PA
REDENÇÃO
A. Venturieri [Embrapa Amazônia Oriental] - PA OURO PRETO DO OESTE - RO
B. Mertens [Cifor]
R. Poccard-Chapuis, M.-G. Piketty [Cirad] V.-Dale [Oak Ridge National Laboratory]
BRASILÉIA AC
M. Pedlowski [Universidade Estadual do Norte Fluminese]
Local support: E.A. Borges, M.C. Rocha da Silva, P.S. Braña Muniz
Local support : L. Ventura dos Santos [APA]
[Pesacre]
Source : ISA, 2000
First preliminary results : cases outline
Ouro Preto do Oeste – RO BR-364, built in 1960-65, paved in 1984, in good general state
Good soils
National colonization programs, massive immigration of colons from the South
History of occupation before the official colonization : e.a. rubber cycles
Significant role of the MST
Brasiléia – AC BR-317, built in 1958, paved in 2002, in very good general state
Relatively good soils and flat topography, favourable to ranching activities
History : Acrean Revolution and old history of occupation by traditional communities
(seringueiros), resistance to the colonization model
Tradition of environmental conservation, influencing political decision process
Border area, project of the “Road to the Pacific”
Guarantã do Norte – MT BR-163, built in 1973-74, paved in 2000-04, in relatively good general state
Area of private colonization
Economic development of the city due to its “end of the road” position
Land tenure system stabilized, area with few land tenure conflicts
Altamira / Uruará – PA BR-230, built in 1970-74, not paved, in bad general state
Area of INCRA planned colonization schemes
Relatively few occupation before the official colonization
Consolidated frontier (old frontier)
São Félix – PA PA-279, built in 1986-92, not paved, in good general state
Road development by the state government following local development of activities
Impact of the gold extraction
Mixed frontier, land speculation and conflicts
Methodology
• Qualitative Comparative Analysis - QCA (Ragin, 1987)
– recent approach and technique for comparing in a systematic way a
limited number of cases („small-N‟)
– via media between traditional quantitative and qualitative approaches
Qlit
N
Qualitative Quantitative
Very small N Very large N abstraction of individual
case study
cases
« case-oriented » ‘Small-N’ « variable-oriented »
deep knowledge of each case
covariance patterns
no quantitative measures
statistical analyses
Methodology (2)
• Why choose QCA ?
– interpretation and limited generalization
• „ Small-N ‟
– complexity of causal relationships
• « multiple conjunctural causation » (and / or)
– controlled observation (quasi-experimental research design)
– combines holistic and analytic approaches
• cases as configurations of conditions
– parsimony
• reduces complexity by highlighting causal regularities
– replicability
Implementation of QCA
• N = 15 to 20 cases
• Variables
– outcome (dependent variable) : high / low deforestation along the road
– potential explanatory conditions (independent variables) : « macro-variables »
• state and age of the road infrastructure
• initial conditions
• local ecological conditions - e.a. soil quality, topography, rainfall, forest type
• local socio-economic conditions - e.a. urban development, population density, markets
• local institutional factors - e.a. protected areas, land tenure system
• local political factors - e.a. local conservation and development policies
• national political context - e.a. incentives, colonization, development programs
• national economic context - e.a. conditions on national markets, inflation
• international economic / political context - e.a. conditions on global markets (soybeans, wood),
conventions on the environment (e.g. Kyoto)
Implementation of QCA (2)
• Expected results
– « minimal equations » : main factors controlling the impacts of roads
– typology of cases interpretation and link with theory
• Difficulties
– dichotomization of variables
– number of potential expanatory conditions macro-variables
– static approach time periods (temporal interpretation, more cases)
• Perspectives
– QCA (binary variables) and Multi-Value QCA (categorical variables)
– application of logistic regression techniques to Small-N conditions
– introduction of cases outside the Brazilian Amazon (Peru and Ecuador)
Policy implications
• Context
– Avança Brasil : e.a. near doubling of paved roads
– alarming predictions : 28 to 42% of the region deforested or heavily
degraded following current development plans and land use trends
(Laurance et al., 2001)
• Guide the implementation of relevant development and
conservation policies :
is it better to intensify roads in already settled areas, or to expand the
road network to primary forest areas in order to make them accessible ?
Any questions
or
comments ?
Collaborations
Study sites Research teams
Altamira PA Indiana University
Uruará PA Michigan State University
Clark University / Cifor / IPAM
Embrapa Amazônia Oriental
Ouro Preto do Oeste RO Oak Ridge National Laboratory
Universidade Estadual do Norte Fluminese
APA (local support)
Brasiléia AC Pesacre (local support)
Université Rennes 2
São Félix PA Cifor / Cirad
Imazon (?)
Redenção PA
Transamazonia network (?)
Marabá PA
GEOMA Project (?)
Equateur - Transamazonia network - Ecuador