Explaining the fragmentation in the Brazilian Amazonian forest

23 Pages·2015·0.9 MB·English

by  Eugenio Y. Arima

This article was downloaded by: [98.180.25.23]
On: 28 April 2015, At: 05:46
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Land Use Science
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/tlus20
Explaining the fragmentation in the
Brazilian Amazonian forest
Eugenio Y. Arimaa, Robert T. Walkerb, Stephen Perzc & Carlos
Souza Jr.d
a Department of Geography and the Environment, The University
of Texas at Austin, Austin, TX, USA
b Center for Latin American Studies, Department of Geography,
University of Florida, Gainesville, FL, USA
c Department of Sociology, University of Florida, Gainesville, FL,
USA
Click for updates d Instituto do Homem e Meio Ambiente da Amazonia, Belém, Pará,
Brazil
Published online: 27 Apr 2015.
To cite this article: Eugenio Y. Arima, Robert T. Walker, Stephen Perz & Carlos Souza Jr. (2015):
Explaining the fragmentation in the Brazilian Amazonian forest, Journal of Land Use Science, DOI:
10.1080/1747423X.2015.1027797
To link to this article: http://dx.doi.org/10.1080/1747423X.2015.1027797
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
5
1
0
2
ril
p
A
8
2
6
4
5:
0
at
]
3
2
5.
2
0.
8
1
8.
9
[
y
b
d
e
d
a
o
nl
w
o
D
JournalofLandUseScience,2015
http://dx.doi.org/10.1080/1747423X.2015.1027797
Explaining the fragmentation in the Brazilian Amazonian forest
Eugenio Y. Arimaa*, Robert T. Walkerb, Stephen Perzc and Carlos Souza Jr.d
aDepartment ofGeographyandtheEnvironment,TheUniversity ofTexasatAustin,Austin,TX,
USA; bCenterforLatinAmerican Studies,DepartmentofGeography,University ofFlorida,
Gainesville,FL,USA;cDepartment ofSociology,UniversityofFlorida,Gainesville, FL,USA;
dInstitutodoHomemeMeio AmbientedaAmazonia,Belém,Pará,Brazil
(Received7 October2013;accepted4March2015)
Althoughvastliteratureexistsonthedriversoftropicaldeforestationanditsecological
5 consequences,lessisknownabouthowpatternsofforestfragmentationemergeinthe
1
0 firstplace.ThepurposeofthispaperistoaddressthisissuefortheBrazilianportionof
2
pril tphaetteArnms,azinocnlubdainsignrebcytanangaullyazr,infgishtbhoeneso,criaadliapl,rodceenssdersiticg,enaenrdatiwvheatofwefivreefesrpetocifaics
A ‘the stem of the rose.’ We argue that forest fragmentation patterns in the Brazilian
28 Amazon are largely determined by the types and arrival times of the agents who
6 engage in land clearing. We also arguethat the patterns manifest in the landscape by
4
5: virtue of road construction and agricultural property formation, which often occur in
] at 0 taannddeombs.eWrveectohnactlufrdaegmbyenptlaatcioinngsoteumrdmisincgusfsrioomnwfoitrhminallecgoalloannizdaitniosntitpurtioojnecatlscoisntmexotrse,
3 consistent with biodiversity conservation than that associated with a laissez-faire
2
5. occupation. However, erosive impacts may be greater in topographically sensitive
2
0. areas.
8
8.1 Keywords: forest fragmentation; road network; Amazonia; land change; process to
9 pattern
[
y
b
d
e
d
a
o
nl
w
1. Introduction
o
D
Forest fragmentation has emerged as an important environmental issue in tropical envir-
onments,giventhatthespatialarrangementandgeometricconfigurationoffragmentscan
compromise ecological processes. For the Amazon Basin, researchers have documented
howforestfragmentationaltersanimalhabitats,compromisesvegetativeregeneration,and
sparks biomass collapse along fragment edges (Aldrich & Hamrick, 1998; Benitez-
Malvido, 1998; Ferreira & Laurance, 1997; Laurance, 1998; Laurance, Delamônica,
Laurance, Vasconcelos, & Lovejoy, 2000; Laurance et al., 1997; Nepstad et al., 2001;
Scariot, 1999). The present article takes up the issue of forest fragmentation in the
Amazon basin, the world’s largest contiguous tropical forest, which continues to disap-
pear, although deforestation rates have come down in recent years. Our goal is to extend
land change science in the direction of explicating the spatial dimensions of land cover
change (Gutman et al., 2004).
Ecological degradation stemming from forest fragmentation concerns us, but our
analytical interest resides in identifying the background factors giving rise to
*Correspondingauthor.Email:[email protected]
©2015Taylor&Francis
2 E.Y. Arima et al.
fragmentation in the first place, specifically the underlying social processes (Arima,
Walker,Souza,Pereira,&Canto,2013;Walkeretal.,2013).Thus,theanalyticalobjective
of the paper is to provide a synthesis of what is presently known about the social
processes generative of fragmentation patterns currently observed in the Amazon Basin,
particularly Brazil.
The key to understanding such generative processes is to gain insight into road-
building. Although roads have long been recognized as an important deforestation driver
(Geist & Lambin, 2002; Nelson & Hellerstein, 1997), the interest here resides in the fact
thatroadnetworkarchitectureplaysacriticalroleinshapingforestfragmentationpatterns
(Arima, Walker, Perz, & Caldas, 2005). Roads are one-dimensional and do not readily
revealthemselveswhenviewedfromafar,asviasatelliteimagery,nordotheyrepresenta
significantbarriertoecologicalprocessesdependentonmobilityorspatialdispersion.The
fragmentation impact of roads comes later, through those who follow and amplify the
spatialsignatureofroadpatternsbyclearingforest,typicallytoestablishagriculturalland
uses (Figure 1) (Arima et al., 2008; Forman, 2003). Thus, an explanation of forest
5 fragmentation resides partly in an explanation of how roads emerge spatially from the
1
0 tabularasaofprimarytropicalforestandpartlyintheformationofagriculturalproperties.
2
ril Theseprocessesoftenoccur intandemandcontemporarypatternsofforestfragmentation
p
A are largely determined by the types and arrival times of the actors who engage in road
8 construction and land clearance.
2
6 Road construction has usually been associated with governmental venture, but private
4
5: citizenshave beenextremely active in this same endeavor in Amazonia. To maintain termi-
0
at nologicalconsistencywiththeliterature,werefertofederal,andalsostate,roadsasofficial,and
3] tothosebuiltbyprivatecitizensasunofficial(Brandão&Souza,2006;Perz,Caldas,Arima,&
2
5. Walker,2007;Perz,Overdevest,Caldas,Walker,&Arima,2007;Perzetal.,2005).
2
0. Our synthesis of fragmentation patterns yields a conceptual framework organized on the
8
1 basis of the agents responsible for forest occupation, and the nature of their actions and
8.
9 interactions.Inpresentingit,webeginthearticlebynotinganumberofempiricalregularities
[
y inspatialpatternsofforestconversionthathavebeenobservedworldwide,andintheAmazon
b
d Basinaswell.Followingthis,weconsideraspectsoftemporalandspatialscalethatenableusto
e
d drawadistinctionbetweenactionstakenbythefederalgovernmenttoopenthebasin,andthe
a
o
nl spatial behaviors of the agents who came to Amazonia as a consequence, in the interest of
w
resourceexploitation.Ourdistinctioninplace,wemoveontoaccountsofhowtheBrazilian
o
D
Figure1. Roadsandlandclearingsdetermine patternsoffragmentation.
Journal of Land Use Science 3
Amazon’s primary fragmentation patterns emerged. Included here are the federal highway
networkandfivedistinctpatternsassociatedwithprivateagentspursuingpersonalobjectives.In
additiontoprocessdescriptions,weprovideestimatesoftheextentsforeachofthesefragmen-
tationpatterns,aspresentlyfoundintheAmazonBasin.Oncewehaveaddressedtheobserved
patterns, we distil our conceptual synthesis, which traces social processes into landscape
patternsonthebasisoftheagentsinvolvedandtheirspatialdecision-making.Wethenconclude
thearticlebyconsideringimplicationsforpolicyandmodelingexercisesinlandchangescience.
2. Methods
The article draws information from the academic literature and from research efforts by
the authors, including field surveys, simulation, and remote sensing. The field surveys
involved hundreds of structured and open-ended interviews of long-time residents, state
officials, agricultural landholders, loggers, and forestry engineers. They were conducted
5 along federal highways BR-364 in Rondônia (2003), BR-163 in Pará and Mato Grosso
1
0 (2004, 2009), and BR-230 in central Pará (2004, 2005, 2010), and in the Terra do Meio
2
ril region of southern Pará (2009, 2011). For BR-230, we also documented the opening of
p 118 roads by private citizens, aside from the federal and state highways inourstudy site.
A
8 Simulations complementing the fieldwork (Arima et al., 2008; Arima et al., 2005, 2013;
2
6 Walker et al., 2013) used a comprehensive GIS data base for the Brazilian Amazon,
4
5: derivedfromLandsatTMandETM+images(Figure2,panelinthecenter)(Arimaetal.,
0
at 2013; Brandão & Souza, 2006). These data were created by manual digitalization of
]
3
2
5.
2
0.
8
1
8.
9
[
y
b
d
e
d
a
o
nl
w
o
D
Figure2. Roadnetworksandpatternsofforestfragmentation.
4 E.Y. Arima et al.
roads, identified by visual inspection of Landsat bands 3 and 5 on a 1:50,000 scale
(Brandão & Souza, 2006). Also used was ‘soil fractional cover,’ which was obtained
throughspectralmixtureanalysis(Souzaetal.,2013).Suchmethodsenhancethecontrast
between exposed soils, which appear as bright pixels, and the surrounding dark vegeta-
tion, enabling visual identification. Overall accuracy of the data approaches 100% for
permanent roads (Brandão & Souza, 2006). The database takes 2003 as a baseline, with
yearly updating to 2012 using cloud-free images. More than 1200 images were used and
almost 100,000 segments digitized.1 The GIS database provides the only comprehensive
coverage of unofficial roads for the Amazon Basin. As such, it enables order-of-magni-
tude calculations for associated road networks.
3. Fragmentation patterns in the Brazilian Amazon
3.1. Overview of fragmentation patterns
5 As a first step toward the synthesis of forest fragmentation processes, we present a
1
20 typology of patterns observed in the Brazilian Amazon. Our approach does not invoke
ril commonlyusedfragmentationmeasuressuchasfractaldimension.Thisisbecauseweare
p
A interested in the actual shapes in question, which are not readily translatable into frag-
8 mentation indices. Helpful to our endeavor is the widespread replication of specific
2
6 geometries associated with frontier penetration. At the global scale, six general patterns
4
5: havebeendiscernedinpreviouswork(Geist&Lambin,2002;Mertens&Lambin,1997),
0
at and Oliveira Filho and Metzger (2006) suggest that three of these are dominant in the
3] Brazilian Amazon, including what they refer to as the ‘fishbone,’ ‘independent settle-
2
5. ment,’andthe‘largeproperty’patterns.Wedonottakeissuewith thisnomenclature, but
2
0. note that it mixes spatial description (e.g., fishbone) with implicated agent (e.g., large
18 property). Consequently, the present article seeks to avoid pattern–agent confusion by
8.
9 presenting a strictly spatial terminology. Each pattern is depicted in Figure 2 (thin lines).
[
y The background image, displayed in false color composite (RGB: bands 7,4,2), is from
b
d NASA’sGeocover2000.Forestsareindarkcolorsanddeforestedareasandexposedsoil/
e
ad rock in light colors.
o
nl Weretainthefishbonelabel (Figure 2,top rightpanel),apt asaspatialdescriptor and
w
with a long tradition in Amazonian research (e.g., Skole & Tucker, 1993), although
o
D
variations have been noted (Batistella, Robeson, & Moran, 2003). For the independent
settlement pattern, we use the term dendritic (Figure 2, bottom right panel), given its
highly irregular, trellis-like structure (Arima et al., 2008; Arima et al., 2005; Chorley &
Haggett, 1967). Finally, for the large property pattern of Oliveira Filho and Metzger
(2006), we adopt the term rectangular (Figure 2, bottom left panel). Reminiscent of the
term, geometric, of Mertens and Lambin (1997), rectangular is more precise in its
reference to the fragments of cleared land often observed in areas of large-scale commer-
cial agriculture, both ranching and mechanized farming.
To the patterns observed by Oliveira Filho and Metzger (2006), we append two that
have come to our attention through fieldwork, including the radial pattern (Figure 2,
upper left panel), and what we refer to as the stem of the rose (Figure 2, mid-left panel).
Radial fragmentation comprises landscape impacts that cluster around urban nodes,
suggestive of the urban pattern of deforestation proposed by Mertens and Lambin
(1997). The stem of the rose combines the visual features of dendritic and fishbone
fragmentation,withevenlyspacedandstubbyspinesthatquicklylosetheirinitialparallel
coherence. As a sixth meta-pattern, we also consider the broad landscape partition
Journal of Land Use Science 5
associated with Brazil’sfederal highways, orofficial roads. The impactofsuch highways
on Amazonian migration and environmental change has been well documented (Mahar,
1988; Moran, 1981; Schmink & Wood, 1984). Less is known about how official roads
distributed unofficial roads across the basin, thereby entraining fragmentation processes.
Consequently, there remains a need to evaluate how fragmentation patterns are related to
each other spatially, and how their various spatial signatures emerged over time.
3.2. Temporal and spatial scale
Having noted a suite of geometric patterns of forest fragmentation, we now turn to the
identification of the specific agents responsible for road construction and subsequent
forest fragmentation, together with their temporal and spatial scales of operation.
Research forBrazil often implicates loggers,smallholders, ranchers, farmers, andminers,
especially forgold. Such individuals seek highly personalized objectives when they clear
forest, whether to liquidate valuable mineral resources like miners, or to engage in long-
5 term agriculture like smallholders and ranchers. Given our basin-scale interest, we aug-
1
20 ment this list with the politicians and bureaucrats whose activities open settlement
ril frontiers in forest regions through the building of highways. Although they do not gain
p
A explicitrewardsfromtheextractionofmineralsoragriculturalactivity,theiractionsfulfill
8 societal objectives, such as economically integrating new spaces of extraction, securing
2
6 national borders, or promoting regional development to alleviate poverty.
4
5: The agents identified pursue their various objectives in specific temporal and spatial
0
at contexts,withthetemporalcontextinvolvingrelativetimingandduration.IntheBrazilian
3] case, politicians and bureaucrats began infrastructure developments prior to those who
2
5. came later to exploit resources; moreover, federal highway construction took a consider-
2
0. able amount of time, about half a decade, and maintenance has persisted over decadal
8
1 periods.Asforresource-orientedactors,theyoftenshowmeaningfulsequencing,aswhen
8.
9 smallholders penetrate forest following the roads built by loggers. Such timing has been
[
y referred to as a logical tandem or sequential occupation (Geist & Lambin, 2001; Walker,
b
d 1987),andisobservedinmanypartsoftheworldbesidesBrazil(Walker,1987;Walker&
e
ad Smith, 1993). In general, road construction by private individuals occurs with surprising
o
nl rapidity,ontheorderofseveralkilometersperdaywhenearthmovingequipmentisused.
w
Unlike the federal system, maintenance is haphazard, and the roads, potentially
o
D
ephemeral.
Thetemporalscaleofoperationarticulateswiththespatialextentofroadbuilding.The
politiciansandbureaucratsresponsibleforBrazil’sfederalandstatesystemsimplemented
anofficialnetworkthattraversesmuchoftheBrazilianportionoftheAmazonbasin,with
some exceptions such as inaccessible terrains to the west of the Purus River Valley in
Amazonas State. The spatial implication is that the first tier of basin fragmentation
emerges from a rectilinear network of federal and state roads that presently covers
50,000 km (Figure 3), and yields a density of 0.011 km−1 (km of road/km2 of area)
(Plano Nacional de Logística e Transportes [PNLT], 2012).2 As for the action realms of
theresourceusers,theseareconsiderablysmaller,giventhelackofcapitalizationandtheir
localized economic interests. Nevertheless, unofficial roadbuilding activities are also
significant, especially if they piggyback on segments of the federal system, as has been
the case with much of the basin’s fishbone fragmentation. The summed length of
unofficial roads across numerous local networks adds up to approximately 460,000 km
(Figure 3), resulting in an overall density of 0.103 km−1, an order of magnitude greater
than the official system (Arima et al., 2008; Arima et al., 2013).
6 E.Y. Arima et al.
5
1
0
2
ril
p
A
8
2
6
4
5:
0
at
]
23 Figure3. Lengthofroadnetworks.
5.
2
0.
18 Tosummarize,theBrazilianAmazon’sofficialandunofficialroadsmanifestimportant
8. scalar characteristics in time and space. Paramount here is the fact that the federal
9
[
y highway system functioned as a temporal and spatial prior for the five patterns of
b
d fragmentation described. The construction and maintenance of official roads represent
e
d long-term political processes spanning presidential administrations and federal planning
a
nlo programs of decadal duration. In contrast, unofficial roads may be ephemeral, even
w seasonal, although profitable agriculture generates a spatial signature that persists as a
o
D discernible fragmentation pattern.
3.3. Agricultural property
Agricultural land use is the process that magnifies the spatial pattern of a particular road
networkintoanactualfragmentationpattern(Arimaetal.,2008;Forman,2003).Asafirst
steptoidentifyingthefragmentationpatternproduced,weassertthatagriculturalproperty
manifestarectangularshape,anassertionthatreflectstheempiricalregularity.Thetheory
of optimal property shape assumes rectangles a priori and focuses on road frontage and
depth dimensions (Colwell & Scheu, 1989). Thus, we suggest that the optimality of
rectangles stems from minimizing demarcation and production costs. As for demarcation
costs, rectangular shapes involve a minimal number of survey points at right angles; this
facilitates boundary identification, thereby reducing transactions costs in land markets.
Because rectangles form convex sets, lines joining points within a property lie within its
boundaries, which reduces production costs, such as for fencing or for the operation of
farm equipment.
Journal of Land Use Science 7
Consequently,itisnotsurprisingthatcommercialinterestsoptforrectangularproper-
ties, an economic preference that underlies rectangular fragmentation. But the same is
true for smallholders, in which case differences in land holdings are manifested in
property size and in the relative dimensions of frontage versus depth. Commercial
interests often show preference for nearly square properties, ranging up to tens of
thousands of hectares. Smallholders occupy much smaller and more numerous parcels,
which exhibit variability in shape as a function of the circumstances of colonization and
land occupation. Where the federal government has promoted settlement, smallholders
occupyelongatedrectangleswithlimitedroadfrontage(e.g.,400m×2500malong BR-
230),and forest clearingmoves fromfront toback, with theresult that afishbonepattern
emergesalongparalleldeforestationstrips,anchoredbythecolonizationroads.3Thus,our
useofthetermrectangularmustnotbeconstruedasrestrictingrectangularpropertiestoa
singlelandscapepattern, sincethisalsodepends onpropertysizeandspatialarrangement
along roads.
5
1
0
2 4. From social process to fragmentation pattern
pril Thusfar,wehavetouchedonthehumanbehaviorsassociatedwithroadconstructionand
A
8 agricultural land use, which underlie most Amazonian forest fragmentation in Brazil. We
6 2 nowconsiderthesocialprocessesandthespatialdecision-makingthatleadtothedifferent
5:4 landscape patterns. Our start point is Brazil’s federal highway network, in particular the
0 segments built to open its Amazonian region. We have referred to roads built by
at
] individual states as also belonging to the category of official roads. However, our
3
2 discussion here addresses the federal system, because state roads formed only a small
5.
2 componentoftheofficialnetworkintheBrazilianAmazonregionformanyyears.Indeed,
0.
8 it was the federal highway network that functioned as the prime distributor offragmenta-
1
8. tion patterns at subbasin scale. Facilitating the network’s distributional role was a sig-
9
[ nificantinterventionbythefederalgovernment,whichinvolvedthetransferofstatelands
y
b in Pará, Mato Grosso, and Rondônia, lying within 100 km on either side of the new
d
e highways, to federal jurisdiction for the purposes of agricultural development (Diário
d
oa Oficial da União [DOU], 1971).4
nl
w
o
D
4.1. The federal network
Until the mid-twentieth century, coastal waters provided the only connection between
Amazonia and the rest of Brazil, and rivers maintained the region’s internal connectivity
(Becker, 1982). Significant infrastructure investment begins with the administration of
President Kubitschek and the building of the Belém–Brasília highway (BR-010), which
was completed in 1960 and paved in 1974. As the military government came into power
in 1964, road construction became part of the regime’s geopolitical strategy to secure
national borders and provide access to resource frontiers (Kleinpenning, 1977; Smith,
1982). The military government began its infrastructure buildup in 1967 with highway
BR-364, which connected Cuiába to Porto Velho by 1974 and Rio Branco by 1975.
Asphalted in 1983, BR-364 now stretches to the Peruvian border, linking northwestern
Braziltothecountry’sindustrialcoreinthesouth.Themilitarygovernmentnexttargeted
central Amazonia, anxious to assert sovereignty over a large, relatively unpopulated
regionwithinitsnationalborders(Hecht,2011).Undertheslogan,‘bringapeoplewithout
land to a land without people,’ massive investments opened the region by the mid-1970s
8 E.Y. Arima et al.
with the completion of three major highways, BR-230 (the Transamazon Highway), BR-
163 (Cuiabá–Santarém), and BR-319 (Porto Velho–Manaus).
BR-230linkedthedrought-strickennortheastofBraziltoItaitubain1972(Sant’Anna,
1998).AlthoughplannersintendedtoreachBenjaminConstantinthetri-borderregionof
Brazil, Peru, and Colombia, construction stopped in 1974 at Lábrea, on the Purus River
(Kleinpenning, 1977; Sant’Anna, 1998). From its start point in the State of Paraíba, the
Transamazon Highway, mostly unpaved through the Amazon, traverses 4223 km. The
second road targeting central Amazonia, BR-163, connects Cuiabá to Santarém over a
1743 km clay track currently being paved. Completed in 1976, BR-163 connects
Santarém, an important town on the Amazon River, to the industrial south via Cuiabá
(Araújo et al., 2008). Finally, BR-319 was built to connect Porto Velho (linked to the
industrial south via BR-364) with Manaus, the largest city in the central Amazon, and a
free trade zone since 1967. Started in 1968 and completed in 1976, BR-319 crosses BR-
230ontheMadeiraRiveratHumaitá(Sant’Anna,1998).Heavyrainshaveminimizedits
developmental impacts, as they have for western segments of BR-230. The period of
5 large-scale Amazonian road construction closed during the military regime with the
1
0 opening of BR-174, linking Manaus to Boa Vista by 1977, and BR-156, linking
2
ril MacapátoOiapoqueby1976(Sant’Anna,1998).Bothwerebuiltpursuanttogeopolitical
p
A concerns, notably to reach and thereby secure borders with Venezuela, Guyana, and
8 French Guiana.5
2
6 In all cases, federal highway design began with the specification of a destination
4
5: determinate route linking population centers or points of strategic interest (Arima et al.,
0
at 2005). Planning bureaucrats also considered the locations of preexisting settlements,
3] which were used by road construction crews as logistical ‘support points’ (pontos de
2
5. apoio). Often, such preexisting settlements reflected Amazonian hydrology (Instituto
0.2 Brasileiro de Geografia e Estatística [IBGE], 1957), as with the east–west route of BR-
8
1 230 that links Marabá, Altamira, and Itaituba (Figure 4). These are all old towns located
8.
9 on or near the fall line of the Brazilian shield and its drop to the Amazon stream course.
[
y Planners paid little heed to topography at local scale, but did avoid the hilly areas of the
b
d Serra do Cachimbo (BR-163) in southern Pará State, as well as extensively inundated
e
d lands in Amazonas State (BR-230). Given the route selection imperatives at play, the
a
o
nl geometry of federal highways in the Brazilian Amazon takes the form of long straight
w
segments that intersect and connect far-flung nodal points. This leads to a very sparse
o
D
network of nearly 17,000 km of nearly continental extent (Figures 3 and 4).
While the federal system connected major population centers within Amazonia, and
betweenAmazoniaandtherestofBrazil,thestatesystemvastlyextendedthatconnection
tosecond-tiercitiesandintoresource-richlocationswithdepositsoftin,gold,andironore
in Pará, Amazonas, and Rondônia States. Although built later, the state system is now
twice the size of the federal system, comprising 32,500 km (Figure 3).
4.2. The distributed fragmentation patterns
TheimpositionofthemainfeaturesoftheFederalNetwork(asmirroredandextendedby
theStateSystem)hadlargelytranspiredbythemidtolate1970s,aperiodcoincidentwith
large colonization projects known as Projetos de Colonização, or PICs, in Pará and
Rondônia State (Cardoso & Müller, 1977; Moran, 1981), as well as with substantial
private initiatives in Mato Grosso and parts of Pará (Jepson, 2006; Schmink & Wood,
1992).Theseinitiativesininfrastructureinvestmentandcolonizationpolicystimulatedthe
expansion of spatially distributed settlement frontiers along the highway corridors, a