Human Pressure Amazon

8/14/2019 Human Pressure Amazon

1/86


2/86


3/86

Each World Resources Institute report represents a timely,scholarly treatment of a subject of public concern. WRI takesresponsibility for choosing the study topic and guaranteeingits authors and researchers freedom of inquiry. It also solicits

GREG MOCK

Editor

HYACINTH BILLINGSPublications Director

MAGGIE POWELLCover Design

RL|2 COMUNICAO E DESIGNLayout

Copyright 2006 World Resources Institute. All rightsreserved.ISBN: 1-56973-605-7Library of Congress Control Number: 2006923217Printed in Belm, Par, Brazil by Grfica & Editora Alves onrecycled paper (Reciclato)Cover Photographs: Cattle (Ritaumaria Pereira); Burningrainforests (Digital Vision); others (Paulo Barreto)Background satellite picture: US Geological Survey.Landsat 7 ETM + Satellite Sensor. 1999

and responds to the guidance of advisory panels and expertreviewers. Unless otherwise stated, however, all theinterpretation and findings set forth in WRI publications arethose of the authors.


4/86

The World Resources Institute and Imazonwould like to acknowledge financial support fromthe Government of the Federal Republic of

Germany through the Ministry of EconomicCooperation, the Dutch Ministry of ForeignAffairs, ABN-AMRO Bank, the Ford Foundation,the William and Flora Hewlett Foundation, andthe Gordon and Betty Moore Foundation.

We thank Paulo Adrio and Andrew Murchie ofGreenpeace Brazil for their initial involvement inthe project and their key role in setting it up.Christoph Thies, from Greenpeace International,also played an important role catalyzing and

implementing the project. The Brazil Ministry ofAgrarian Development (Ministrio doDesenvolvimento Agrrio) kindly provided data onland reform projects, while Andrew Murchiefurnished the map of the location of forestoperations used in the analysis.

The authors also wish to thank thosecolleagues who provided valuable reviewcomments: Mark Cochrane, South Dakota State

University; Jos Maria Cardoso da Silva,Conservation International-Brazil; Ernesto

A C K N O W L E D G M E N T S

Alvarado, Washington State University; TomLovejoy, the H. John Heinz III Center for Science,Economics and The Environment; and Andrew

Murchie and Paulo Adario from GreenpeaceBrazil.

Many colleagues within WRI and Imazon helpedus with this study. Within WRI, Dirk Bryant, MartaMiranda, Janice Wiles and Ralph Ridder wereinvolved in early stages of the project. SusanMinnemeyer, Pierre Methot, Janet Ranganathan,David Jhirad, Lars Laestadius, Ralph Ridder andLindsey Fransen provided valuable comments andreviews. Special thanks are due to Isabel Munilla

and Jonathan Lash for their support, guidance,and assistance throughout various stages of theproject. Hyacinth Billings and Maggie Powellhelped getting this report into print; we thank PaulMackie, Phil Angell, and Nate Kommers for theirassistance with outreach, and Stephen Adam, GayleCoolidge and Josh Neckes for their help in thereview and production processes. Within Imazon,Adalberto Verissimo provided invaluable commentsand Mrcio Sales assisted with the analysis of the

correlation between roads and human pressure inprotected areas.


5/86


6/86

FOREWORD . . . . . . . . . . . . . . . . . . . . . . . . . . . 09EXECUTIVE SUMMARY . . . . . . . . . . . . . . . . . 11

1.INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . 19

2.BACKGROUND AND TRENDS INHUMAN OCCUPATION IN THEBRAZILIAN AMAZON . . . . . . . . . . . . . . . . . . . 25

Deforestation . . . . . . . . . . . . . . . . . . . . . . . . . 25Urban Zones . . . . . . . . . . . . . . . . . . . . . . . . . 29Agrarian Reform Settlements . . . . . . . . . . . 29

Forest Fires (Fire Zones) . . . . . . . . . . . . . . . 31Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Roads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

3.MAPPING HUMAN PRESSURE IN THEBRAZILIAN AMAZON . . . . . . . . . . . . . . . . . . . 41

How Human Pressure was Mapped . . . . . . 41Human Pressure in Protected Areas . . . . . . 43How the Relationship between Roads

and Human Pressure was Analyzed . . . . . . 43

4.HUMAN PRESSURE IN THE BRAZILIANAMAZON: RESULTS AND DISCUSSION . 45

C O N T E N T S

Areas under Pressure fromHuman Settlement . . . . . . . . . . . . . . . . . . . . 48

Areas under Incipient Human Pressure . . 495.HUMAN PRESSURE ANDPROTECTED AREAS: RESULTSAND DISCUSSION . . . . . . . . . . . . . . . . . . . . . 55

Human Pressure in ExistingProtected Areas . . . . . . . . . . . . . . . . . . . . . . . 55Human Pressure in Priority Areas forConservation . . . . . . . . . . . . . . . . . . . . . . . . . 60Risks and Opportunities for the Creation

of Conservation Areas . . . . . . . . . . . . . . . . . 64

6.CONCLUSIONS ANDRECOMMENDATIONS . . . . . . . . . . . . . . . . . 67

NOTES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

APPENDIX 1: PROTECTEDAREAS IN BRAZIL . . . . . . . . . . . . . . . . . . . . . . 71

APPENDIX 2: ECONOMICACCESSIBILITY OF LOGGING . . . . . . . . . 73

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . 77

HUMAN PRESSURE ON THE BRAZILIAN AMAZON FORESTS

v


7/86

Figure 1. Vegetation and Deforestation inthe Brazilian Amazon as of 2001 . . . . . . . . . . . 27Figure 2. Distribution of TransportationNetwork, Timber Processing Capacity,and Cattle Processing Capacity . . . . . . . . . . . . 28Figure 3. Population in the Legal AmazonBetween 1950 and 2000 . . . . . . . . . . . . . . . . . . . 29Figure 4. Municipal Seats andUrban Zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Figure 5. Area of Agrarian ReformSettlements and Number of Projects in the

Legal Amazon Between 1995 and 2002 . . . . . . 31Figure 6. Agrarian Reform SettlementsEstablished as of 2002 . . . . . . . . . . . . . . . . . . . . 32Figure 7. Fire Zones, 1996-2002 . . . . . . . . . . . 34Figure 8. Area Allocated for Mining orMining Exploration as of 1998 . . . . . . . . . . . . . 35Figure 9. Location of loggingpermits in 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . 37Figure 10. Land Cover Classificationin Eastern Amazonia . . . . . . . . . . . . . . . . . . . . . 38Figure 11. Human Pressure in the

Brazilian Amazon - All Indicators . . . . . . . . . . 46Figure 12. Two Types of Human Pressure . . . 47

Figure 13. Cumulative Area (Percent) ofHuman Pressure in Relation to Distancefrom Official Roads . . . . . . . . . . . . . . . . . . . . . . 52Figure 14. Types of Access to Fire Zonesin South-Central Par and NorthernMato Grosso . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53Figure 15. Unofficial Roads, HumanPressure, and Indigenous Lands . . . . . . . . . . . 53Figure 16. Distribution of Human Pressurein Existing Protected Areas and inUn-Protected Areas . . . . . . . . . . . . . . . . . . . . . . 56

Figure 17. Human Pressure inProtected Areas . . . . . . . . . . . . . . . . . . . . . . . . . . 57Figure 18. Distribution of Land Cover andHuman Pressure in Un-protected Areasand by Categories of Protected Areas . . . . . . . 58Figure 19. Distribution of Land Cover andDisaggregated Human Pressure inUn-protected Areas and by Categories ofProtected Areas . . . . . . . . . . . . . . . . . . . . . . . . . . 59Figure 20. Human Pressure in PriorityAreas for Conservation . . . . . . . . . . . . . . . . . . . 61

Figure 21. Human Pressure in AreasSuitable for Public Production Forests . . . . . . 63


vi

Box 1. Deforestation andBiodiversity Loss . . . . . . . . . . . . . . . . . . . . . . . . 19Box 2. Protected Areas

in Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Table 1. Human Pressure in theBrazilian Amazon . . . . . . . . . . . . . . . . . . . . . . . . 45

Table 2. Overlap Between Location ofLogging Operations and Other Indicatorsof Human Pressure . . . . . . . . . . . . . . . . . . . . . . 51

Box 3.The Brazilian Amazon, theAmazon Basin, and the Legal Amazon . . . . . . 23Box 4. Mapping Unofficial Roads in

the Midlands of Par . . . . . . . . . . . . . . . . . . . . . 40


8/86

ARPA Protected Areas in the Amazon ProgramDNPM National Department for Mineral ProductionEMBRAPA/CPATU Brazilian Agricultural Research Institute/Center for Agricultural

Research in the Humid TropicsFAO Food and Agriculture OrganizationGFW Global Forest WatchIBAMA Brazilian Institute for Environment and Renewable Natural ResourcesIBGE Brazilian Institute for Geography and StatisticsIMAZON Amazon Institute of People and the EnvironmentINCRA Brazilian Institute for Colonization and Agrarian ReformINESC Institute for Socio-Economic StudiesINPE Brazilian Institute for Space ResearchISA Socio-Environmental InstituteMDA Brazilian Agrarian Development Ministry

MMA Brazilian Ministry of EnvironmentNOAA National Oceanic and Atmospheric AdministrationWRI World Resources InstituteWWF World Wildlife Fund


vii

L I S T O F A C R O N Y M S


9/86


10/86


9

F O R E W O R D

Brazil has the largest expanse of tropicalrainforest in the world, comprising some 40percent of the worlds remaining tropical forest

cover. As such, it represents both an extraordinaryresource for the people of Brazil, and also aninvaluable asset for the people of the world.

Millions of my people depend, directly orindirectly, on the Amazon for their livelihood:farmers sell crops at home and into the worldmarkets; the forest sector accounts for almost 8percent of our countrys annual wealth. And thegreat Amazon basin sustains millions ofindigenous people who depend on the richness of

this place to provide for their every need.

That the Southern hemisphere experienced itsfirst hurricane in recorded history last year makesit even more critical that we recognize the value ofsuch ecosystems as the Amazon Basin and thecritical service they provide to the globalcommunity. They also serve as regulators ofclimate, vaults of biodiversity, and great naturalcleansing engines.

As we become increasingly aware of theimportant role that large ecosystems play in ourcurrent and future lives, we begin to understandjust how important it is that we mange themwisely and with a very long-term perspective.

This report provides us with an extraordinaryset of tools to use in that effort. The scope anddetail of these maps are vital if we are to make the

necessary and unavoidable choices in the future tobalance the needs of our people in Brazil with thelegitimate needs of the worlds people increasinglyimpacted by global environmental change.

We know that the Amazon is subject tosignificant human pressure, and we can seewhere it is, its nature, and its impact. What wecan also see, however, is that humansettlementhuman incursions into the Amazontend to take on a life of their own. Approved and

defined development becomes the trigger forwhat this report terms incipient development,that is, emerging pressure on the Amazon that isnot planned. The maps here are unequivocalabout that.

One of the actions of which I am most proud asPresident of Brazil was the establishment of theAmazon Region Protected Areas. With ourpartners, this program recognizes the

international environmental reach of the Amazonand it also accommodates the legitimateaspirations of Brazils most impoverished.

What these incomparable set of maps, and theaccompanying report show, however, is that while


11/86


10

there is more than sufficient unspoiled area in theAmazon to meet our preservation goals, we mustbe vigilant to the pressure on the Basin thatradiates out from settlements, and, as importantly,the impact of isolated development that disruptsintact ecosystems and does damage in ways wehave yet to fully understand. These maps paint astark and compelling picture.

Brazil is acutely aware of the dual responsibilitythat nature, geography, colonial exploration, and

statecraft have placed upon it. It does not shirk itsunique obligation to the world community that thespecial circumstance of the Amazon Basindemands. Nor does Brazil intend to shirk its dutyto its people to provide them with the opportunityfor better, productive lives.

Balancing those enormous demands requireswise choices and the tools and information toinform those choices. That is the inestimablevalue of this report.

Fernando Henrique CardosoFormer President of the Federative Republic of Brazil


12/86


11

In 2002, approximately 47 percent of theBrazilian Amazon was under some type of humanpressure including deforested areas, urban zones,agrarian reform settlements, areas allocated formining and mining exploration as well as areasunder pressure as indicated by incidence of fire.

Our analysis suggests that in 2002 thereremained enough forest area without evidence of

human pressure to fulfill the governments goal toexpand the protected areas system. This includesabout one million km2 of land considered priorityfor establishing new protected areas and publicproduction forests. However, we can assume thatthe pressure has continued to increase in theregion, so the government must act quickly toimplement their protected areas goal.

The Brazilian Amazon harbors about one-thirdof the worlds tropical forests, an area coveringsome 4.1 million square kilometers. However,land-use conversion in the Brazilian Amazon istriggering forest loss and degradation and rapidly

changing the regional landscape. FAO data revealthat Brazil accounted for approximately 42 percentof global net forest loss from 2000 to 2005; most ofthis deforestation occurred in the BrazilianAmazon.

Establishing new protected areas in lands free ofpressure will help prevent land use conflicts.Areas of incipient human pressure are stillvaluable for conservationdue to relatively lowintensity of usebut the political costs ofestablishing protected areas in these zones willbe higher given the interests already in place.

About 80 percent of the total area

deforested is located within 30 km of an officialroad. However, about half of the area of older firezones(1996-1999)and two-thirds of morerecent fire zones(2000-2002) are located fartherthan 30 km from roads. There is a need to betteraccount for this environmental impact in theplanning, building, and maintenance oftransportation corridors.

Key Findings

E X E C U T I V E S U M M A R Y

In response to public demand for forestconservation many stakeholders are attemptingto reconcile economic development andconservation through initiatives that includeforestry regulation, enforcement of

environmental legislation, and the creation ofprotected areas. Given the rapid expansion ofactivities such as cattle ranching, agriculture,and logging, these initiatives must quickly targetpriority areas to be successful. This demands


13/86


12

accurate and detailed information on the currentstate of Amazon forests and the pressures theyface. Yet such information remains elusive.Despite advances such as satellite imaging, ourunderstanding of the extent and degree ofhuman activities in the Brazilian Amazon is onlypartial. Even deforestation (forest clear cutting)is not fully understood. Up to 1997, the BrazilianSpace Agency (INPE) mapped deforested areasgreater than 6.5 ha. Since then, INPE hasimproved mapping technique but it still misses

deforested areas smaller than 3 ha.

Identifying small deforestation plots andother indicators of incipient human activities inforests could flag areas at risk of increaseddeforestation and forest degradation.Pinpointing these areas at risk would providestrategic guidance for conservation andsustainable development in the region. Butdespite this potential, no comprehensiveanalysis that integrates such spatial data with

other standard measures of forest condition yetexists to help conservation and developmentplanners understand the true extent of humanactivities in the region.

This report seeks to help fill this gap. Itcompiles and integrates geospatial information onvarious indicators to present a picture, roughly asof 2002, of the human pressure on forests in theBrazilian Amazon.Human pressure, for this

report, is defined broadly as the presence ofhuman activities that lead to forest loss anddegradation.

The report distinguishes two major types ofareas under human pressure:

Areas under pressure from humansettlements. In these areas, human presence isfully established, settlements are permanent, andland use tends to be more intensive.Environmental impacts in these areas are higherthan in occupation frontiers because of greaterforest fragmentation as well as urban andindustrial activities. Three indicators wereanalyzed in this category:deforested areas, urban

zones, and agrarian reform settlements.

Areas subjected to incipient humanpressure. In these areas human presence may betemporary, but in some cases people will settle inthe future and influence the forest condition (forexample, fragmenting the forest ecosystem).Logging, wildcat (part-time) mining, hunting,harvesting of non-timber forest products, andshifting cultivation are some of the humanactivities that occur in these areas. Two indicatorswere analyzed in this category:fire zones andareas allocated for mining and mining exploration.

For the purposes of this analysis,fire zones aredefined as areas of human activity associated withthe incidence of forest fires.

At the time of the analysis, no comprehensiveinformation onroads and logging in theBrazilian Amazon region was available. Thus,the map of human pressure we produced didnot factor in these two important indicators ofhuman activity. However, understanding the

crucial role of these two factors in forest impact,we did examine the relationship among humanpressure, roads, and logging using availableinformation. Finally, we also examined therelationship between human pressure and

protected areas.


14/86


13

FINDINGS

In 2002, approximately 47 percent of theBrazilian Amazon was under some type of humanpressure, either as areas under pressure fromhuman settlements (19 percent) or areassubjected to incipient human pressure (28percent).

Areas under pressure from human settlementwere found primarily along official roads in the

so-called arc of deforestation, comprising theeastern and southern edges of the forests in thestates of Rondnia, Mato Grosso, and Par. Othersignificant locations under human pressure werealong the Trans-Amazon highway in the State ofPar, along the Amazon River between Manausand Belm, along the Cuiab-Santarm highwaynear the city of Santarm, and around the mainurban centers in the states of Roraima andAmap.

Areas showing incipient human pressure weregenerally clustered and adjacent to areas of humansettlements, indicating frontier expansion. Thiswas especially true in the states of Par, MatoGrosso, and Rondnia. There were, however,isolated areas of incipient human pressure alongnavigable rivers throughout the region. Such areasappeared to be associated primarily withtraditionalmestizo communities and indigenouspopulations.

PRESSURE FROM HUMAN SETTLEMENTS

Deforested Areas. In 2001, deforested areascovered 11 percent of the Brazilian Amazon.

Cattle ranching is the predominant land use indeforested areas throughout the region(Schneider et al 2002). Between 1990 and 2003,the cattle herd in the Legal Brazilian Amazonincreased from 26.6 million to 64 million head,representing a 140 percent increase (IBGE 2005).

Urban Zones. The area under pressure fromurban zones covered 6 percent of the BrazilianAmazon. Urban zones were identified as the areawithin a 20-km radius around the regions 450

municipal seats as of 1997 (the most updateddataset available at the time of the analysis) (IBGE1999). The 20-km radius was based on fieldobservations, and was intended to capture areasunder pressure from urban populations, such assmall-scale clearings used for periurbanagriculture, spontaneous settlements, intensiveextraction of forest products, waste deposits, andrelease of untreated sewage.

Agrarian Reform Settlements. The extent of

Agrarian Reform Settlements (lands granted bythe federal government to landless people)established as of 2002 (INCRA 2002) coverednearly 5 percent of the Brazilian Amazon. Of thisarea, about half is made up of forested areas thatdo not overlap with any other indicator of humanpressure.

INCIPIENT HUMAN PRESSURE

Fire Zones. Approximately 28 percent of theBrazilian Amazon was subjected to incipient humanpressure associated with fire activity. Fire zones aredefined as the 10-km radius around a forest fire, asdetected by satellite between 1996 and 2002.


15/86


14

Human pressure may vary within fire zones.Two-thirds of fire zones are concentrated neardeforested or urban zones; forests in these areashave likely been subjected to intensive pressuressuch as logging. The remaining third is found inmore isolated locations, indicating fires caused byshifting agriculture carried out by traditional

mestizo communities or indigenous populations.Forests in these areas may be subjected to lessintensive pressures such as hunting, harvesting ofnon-timber forest products, and selective logging.

There are overlaps betweenfire zonesandother indicators of human activity:

With areas affected by selective logging,including the location of half of the selectivelogging permits issued by the BrazilianGovernment;

With the location of informal roads(identified from satellite imagery) in thenorthern portion of Mato Grosso and

South-Central Par.

Between 2000 and 2002 the number of annualforest fires nearly doubled from 22,000 per year toalmost 43,000 per year, showing an acceleration ofincipient human activity.

Areas Licensed for Mineral Exploration. In1998, the total area legally allocated for mineralexploration and mineral reserves was

approximately 2 percent of the Brazilian Amazon.More than half of the allocated area overlappedwith other indicators of human pressure.However, should areas licensed for mineralexploration become economically viable in thefuture, improved access and services could spark

rapid in-migration and deforestation. Goldmining, for instance, has been an importantcatalyst of colonization in the Tapajs MiningReserve in western Par.

LOGGING

The total area of selectively logged forests inthe Amazon is unknown, although estimatesindicate this activity may affect 10,000-20,000 km2

of forest per year in the Brazilian Amazon(Nepstad et al. 1999; Matricardi et al. 2001;Cochrane 2000; Asner et al. 2005). Some of theseforests are converted to agricultural and pastureland soon after timber is harvested, while otherareas remain as logged forest. Evidence suggeststhat most logged forests are within the areas ofhuman pressure identified in this report.However, analysis is needed to fullyand accuratelymap the extent of logging.

A substantial share of the timber harvested inthe Brazilian Amazonestimated at 47 percent in2001 and 43 percent in 2004is thought to beillegal (Lentini et al. 2005). These figuresprobably represent an underestimate of illegallogging, since numerous licensed loggers fail toimplement forest management plans or harvestillegally in public unclaimed lands. Asconventionally practiced, logging causessubstantial damage to the forest, especially when

associated with wildfires. Some companies haveadopted best practices and have obtained greencertification. Nevertheless, there are no recentindependent evaluations of approved loggingoutside certified operations. There is a need tocombine fieldwork with interpretation of satellite


16/86


15

imagery to systematically monitor the impacts oflogging throughout the Brazilian Amazon.

ROADS

About 80 percent of the total area deforested islocated within 30 km of an official road. However,about half of the area of older fire zones(1996-1999)and two-thirds of more recent fire zones(2000-2002) are located farther than 30 km from roads.

Human activity in distant areas is possiblegiven access by rivers and a growing network ofunofficial roads opened by loggers, ranchers, andminers. In southern Par, for instance, about17,000 km of roads were built between 1985 and2001, 60 percent of them on unclaimed publicland. Protected areas seem to slow the advance ofunofficial roads; average growth rates forunofficial roads inside protected areas are threetimes lower than those outside protected areas

(Souza et al. 2004).

HUMAN PRESSURE ANDPROTECTED AREAS

Some 28 percent of protected areas have beensubjected to human pressure. This is significantlysmaller than the percentage of forest areasshowing human pressure outside protected areas,

which totals 59 percent. As proximity to roadsincreases (< 25 km), there is a significantincrease of deforestation and fires withinprotected areas. Increasing transportationinfrastructure without a corresponding capacityfor enforcement is likely to result in greater

human pressure on protected areas. Betterinfrastructure may also increase demands toshrink existing protected areas to benefit theexpansion of agribusiness, as the StateGovernment in Mato Grosso approved in 2003.

Human Pressure in Non-Protected PriorityAreas for Conservation

About 48 percent of the non-protected areasidentified as a priority for biodiversity

conservation (Capobianco et al. 2001) showevidence of human pressure. Lands underpressure from human settlements account for 18percent and those areas subjected to incipienthuman pressure account for 30 percent. Most ofthe area under human pressure is in the easternand the southern Brazilian Amazon, and along thelargest rivers such as the Lower and MiddleAmazon and the Upper Rio Negro.

Human Pressure in Potential Public

Production Forests

Nearly 30 percent of the 1.5 million km2

identified in 1999 (Verissimo et al. 2000) as havingthe potential to become public production forestsshow some type of human pressure. Furthermore,the majority of these areas showing humanpressure overlap with the areas identified to beeconomically accessible for logging (Verissimo etal. 1999). Economic accessibility to logging would

be beneficial to promote sustainable use of forestsif appropriate concession regulation andenforcement capacity were in place. However,insufficient regulation and ineffectiveenforcement has spurred illegal and predatoryoccupation of some existing National Forests.


17/86


16

Risks and Opportunities for the Creationof Protected Areas

About one million km2 of land consideredpriority for establishing new protected areas andpublic production forests have not yet been affectedby human pressure, according to our analysis. Thisis enough land for the federal and stategovernments to achieve their stated goals, whichare to expand the protected areas system to 270,000km2 of conservation-oriented lands by 2009 and

395,000 km2 of public production forests by 2010.Some areas remain valuable for conservation

purposes even though they are subject to incipienthuman pressure. However, in some instances, thepresence of settlers, loggers, and gold miners ishindering the establishment of protected areas infavor of other more popular alternatives, such asagrarian reform settlements or the establishmentof titled land. In fact, the Brazilian Congressrecently ratified legislation foregoing a biddingprocess for titling small land holdings (less than 5

km2) on public lands in the Brazilian Amazonoccupied before December 2004, thwarting anypossibility for these lands to be considered for theestablishment of protected areas. The Ministry ofLand Reform expects to grant titles for more than20,000 km2 of public lands, benefiting 150,000families (MDA 2005). The continued and rapidexpansion of human pressure requires rapidaction by governments to create protected areaseven before incipient human pressure occurs.

Responding to a national demand forconservation and encouraged by the lendingpolicies of Multilateral Development Banks, somestate governments have supported the creation ofpublic production forests and reserves forsustainable development. This has occurred

despite local opposition to strictly protected areasand indigenous territories. Therefore, societaldemand and rapid government action can work forthe protection of priority areas for conservation.

POLICY IMPLICATIONS

As our analysis shows, mapping andmonitoring deforestation alone is not sufficient tounderstand the full range of human pressure on

forests in the Brazilian Amazon. A morecomprehensive analysis of pressures from humansettlements, logging, roads, fires, and othersources is required. The findings presented herehave several implications for public policies:

Vast areas in the eastern and central portions ofthe Brazilian Amazon show evidence of humanpressure, especially in the form of fires.Nonetheless, the area that does not show evidenceof human pressure is large enough for the federalgovernment to meet its goal to expand and

consolidate the protected areas system by 2010.However, the opportunities are diminishing.Human pressures are expanding rapidly, asindicated by trends in deforestation, cattleranching, human population growth, and others.This expansion requires rapid action. Recentfederal legislation allowing temporary limitation ofland use in areas of interest for conservation couldbe applied in areas identified in this report toestablish new protected areas. This has already

occurred in an 82,000-km

2

area in western Par inwhich 68,000 km2 of protected areas were createdrecently.

Roads are correlated with human pressure. Inorder to protect priority areas for conservation,planned investments in road infrastructure should


18/86


17

be accompanied by the creation of protected areasin identified priority areas for conservation. Inaddition, efforts must be made to shield existingprotected areas within the reach of the new orimproved infrastructure. In this regard, the effortby the Brazilian government to create protectedareas before paving the Cuiab-Santarmhighway (BR-163) is commendable. Thisapproach should be applied along other proposedroad-paving routes.

As illustrated by the case in the State of Mato

Grosso, when human pressure increases due togreater economic opportunities, thegovernments commitment to protection maywaver. This may even lead, as it did in MatoGrosso, to the reduction of the size of protectedareas. Further analysis and policy debate to fostera long-term government commitment toprotected areas is needed.

The evaluation of human pressure in theBrazilian Amazon presented in this report is aninitial effort that will benefit from further

refinement. Data limitations prevent full and

precise analysis, and more detailed information isneeded. For instance, a comprehensive map of theroad network in the region does not exist. Imazonis currently addressing this limitation by digitizingfrom satellite imagery visible roads in theBrazilian Amazon. Likewise, an accurate andcomplete map of logged forests as well as otherforms of forest degradation (such as burnedforests) is unavailable. Existing remote sensingtechniques, complemented with ground-truthingat strategic sites, will allow mapping of these

features. Investment in these types of research iscrucial to bring a clearer understanding of theextent and intensity of human pressures in theforests of the Brazilian Amazon.

Despite these limitations our analysis providesa more complete picture than formerly available ofthe dimensions of human pressures in theBrazilian Amazon and the diverse forms thesepressures take. As such, it can help guide strategicactions to improve forest conservation until betterinformation becomes available.


19/86


20/86


19

The Brazilian Amazon harbors about one-third of the worlds remaining tropical forests inan area covering some 4.1 million square

kilometers. However, land-use conversion isoccurring at unprecedented scales and in acomplex manner. The mean annualdeforestation rate from 2000 to 2005 (22,392km2 per year) was 18 percent higher than in theprevious five years (19,018 km2 per year). FAOdata (FAO 2005) reveal that Brazil accounted forapproximately 42 percent of global net forestloss from 2000 to 2005; most of thisdeforestation occurred in the Brazilian Amazon.As in other humid tropical forest regions

worldwide, the consequences of this rapidchange include losses of biological and culturaldiversity, changes in the regional and potentiallythe global climate, and social conflicts (see Box 1).Projections indicate that the forces drivingdeforestation and forest degradationsuch asdemand for timber and agricultural productswill continue to grow in the next decade (Zhu etal. 1998; USDA 2005; OECD/FAO 2005). Forestsloss is likely to increase if current trends prevail.

In response to mounting public concern overdeforestation and forest degradation in the region,many stakeholders are attempting to reconciledevelopment and conservation through initiativesthat include the enforcement of environmental and

1INTRODUCTION

An estimated 10-20 percent of all knownspecies live in Brazil (Capobianco et al. 2001;Guimaraes Vieira et al. 2005). Deforestation hasimportant implications for this wealth ofbiodiversity, since many of these species inhabitthe forests of the Amazon Basin. The number ofindividual organisms affected by deforestation inBrazil is unknown, but is bound to be high, giventhe density of organisms and species per hectareof Amazon forest. One recent estimate suggeststhat as many as 50 million birds could have beendirectly affected on the 26,000 km2 deforested inBrazil between 2003 and 2004 (Guimaraes Vieiraet al. 2005). The number of primates affected inthis period was estimated at 2 million.

Biodiversity is not the only casualty ofdeforestation. Other ecosystem services are alsoaffected. Soil erosion, nutrient depletion, loss of

watershed regulation functions, and emission ofgreenhouse gases are some of the more damagingecosystem costs of forest clearing and degradation(Fearnside 2005).

BOX 1 DEFORESTATIONAND BIODIVERSITY LOSS


21/86


20

forestry legislation, the creation of protected areas,and proposals to regulate the use of public forests.Specific government action along these lines includes:

Increase in legal reserves: by law, privatelandowners can only clear a proportion oftheir holdings; the portions that remainforested are known as legal reserves. In 1996,the Federal government increased the legalreserves from 50 percent of the landholdingsto 80 percent. This means that private

landowners can only clear up to 20 percent oftheir lands. The Brazilian Congress approvedthis measure in 2001.

Higher fines for environmental crimes:in 1998, the Brazilian Congress passedlegislation that raised fines for environmentalcrimes (including illegal logging,deforestation and fires) from a maximum ofUS$2,200 to a maximum of US$ 22 millionper event.1 The fine for illegal deforestationis nearly US$700 per ha (0.01 km2).

Creation of protected areas: protectedareas (see Box 2) remain the central featureof efforts to conserve the regions biodiversityand promote sustainable use of naturalresources. By May 2004 about 32 percent ofthe Brazilian Amazon was contained in 427protected areas; a quarter of these wereindigenous lands (Capobianco et al. 2001;Viana and Valle 2003; ISA 2004). In March2000, the federal government launched the

Protected Areas in the Amazon Program(ARPA), aiming to create new protectedareas. The programs goals includeestablishing 270,000 km2 in strictly protectedareas and 90,000 km2 in areas for sustainabledevelopment by 2009. The federal

government also launched in 2000 theNational Forest Program, which proposes toexpand the area of National or State Forests(public production forests) from 85,000 km2

to 500,000 km2 by 2010 (MMA 2000a). Publicproduction forests are to be sustainablymanaged for the production of forest andnon-forest products. Under the forestconcession legislation approved in February2006 these areas would be leased forsustainable use through public bids.

State governments have also establishedorcommitted to establishnew protected areas, aspart of projects financed by multilateraldevelopment banks. Such is the case in the stateof Par in a project financed by the World Bank, orin the state of Acre, in a project financed by theInter-American Development Bank.

Given the rapid expansion of activities such ascattle ranching, agriculture, and logging, these

initiatives must quickly target priority areas to besuccessful. This demands accurate and detailedinformation on the current state of the BrazilianAmazon forests and the pressures they face. Yetsuch information remains elusive. Despitetechnological advances, our understanding of theextent and degree of human activities in theregion is only partial. Even deforestation (forestclear-cutting) is not fully understood. Up to 1997,the Brazilian Space Agency (INPE) mapped

deforested areas greater than 6.5 ha. Since then,INPE has improved mapping techniquesusing adigital systembut it still misses deforested areassmaller than 3 ha. Therefore, small scale shiftingagriculture and incipient occupation may not becaptured until they reach larger sizes.


22/86


21

The Brazilian protected areas system(Sistema Nacional de Unidades deConservao da Natureza, or SNUC)encompasses federal, state and municipal

BOX 2 PROTECTED AREAS IN BRAZIL

PROTECTED AREAS IN THE BRAZILIAN AMAZON

protected areas (see map below). Appendix 1show how SNUC categories relate to the IUCN(International Union for Conservation ofNature) categories.

Source: Capobianco et al. 2001.


23/86


22

Identifying these small deforestation plots andother indicators of incipient human activities inforested areas could flag areas at risk of increaseddeforestation and forest degradation. Pinpointingthese areas at risk would provide strategicguidance for conservation and sustainabledevelopment in the region. But despite thispotential, no comprehensive analysis thatintegrates such spatial data with other standardmeasures of forest condition yet exists to helpconservation and development planners

understand the true extent of human activities inthe region.

This report seeks to help fill this gap. Itcompiles and integrates geographicalinformation on major human pressures in theBrazilian Amazon (see Box 3). Human

pressure, for the purposes of this study, isdefined broadly as the presence of humanactivities that lead to forest loss anddegradation.

This report distinguishes two major classes ofareas under human pressure:

Areas under pressure from humansettlements. In these areas humanpresence is fully established, settlementsare permanent, and land use tends to bemore intensive than in occupationfrontiers. Environmental impacts tend to

be higher because of greater ecosystemfragmentation as well as urban andindustrial activities.

Three indicators were analyzed in thiscategory: deforested areas,urban zones, andagrarian reform settlements.

Areas subjected to incipient humanpressure. Incipient human pressure isassociated with low density subsistence humansettlements and initial occupation connected tomarket demand. In areas subjected to incipienthuman pressure, human presence may betemporary, but in some cases people will settle

in the near future and influence theenvironment (for example, fragmenting forestsand habitats). Logging, wildcat (part-time)gold mining, hunting, and harvesting of non-timber forest products are some of the humanactivities that occur in these areas.

Two indicators were examined in this category:areas associated with forest fires orfire zones(which are also relevant to deforested plotssmaller than 6.5 ha); andareas allocated for

mining and mining exploration.

At the time of the analysis, no comprehensiveinformation onroads and logging in the BrazilianAmazon was available. Thus, the map of humanpressure did not factor in these two importantindicators of human activities. However,understanding the crucial role of these two factorsin forest impact, we did examine the relationshipbetween human pressure, roads, and logging

using available information. Finally, we alsoexamined the relationship between humanpressure andprotected areas.


24/86


23

The Brazilian Amazon overlaps with two othergeographic areas: the Amazon Basin and the LegalAmazon. The Amazon Basin extends over 6.8million square km through Venezuela, Colombia,Ecuador, Peru, Bolivia, Brazil, French Guiana,Guyana, and Suriname (Goulding et al. 2003).Sixty percent of the Amazon Basin lie withinBrazils boundaries, and this portion is known as

the Brazilian Amazon (shaded area within the blackline in the figure below). The Legal Amazon (blackline) contains more than 50 percent of Brazilsterritory and is an administrative unit that

BOX 3 THE BRAZILIAN AMAZON, THE AMAZON BASIN, AND THE LEGAL AMAZON

encompasses the states of Acre, Amazonas,Roraima, Amap, Par, Rondnia, Mato Grosso,Tocantins, and Maranho. Portions of the states ofMaranho, Tocantins, and Mato Grosso are outsideof the Amazon Basin. While 86 percent of theoriginal vegetation of the Brazilian Amazonconsists of dense forests, it also includes openforests in the Amazon-Cerrado transition and

savanna-shrub vegetation (Campinaranas) in theupper Rio Negro (Capobianco et al. 2001). Finally,most of the existing official statistics available arefor the Legal Amazon.

THE BRAZILIAN AMAZON, THE AMAZON BASIN, AND THE LEGAL AMAZON

Source: Dinerstein et al. 1995.


25/86


26/86


25

For centuries, human settlement in the

Brazilian Amazon occurred along the principalnavigable rivers of the region. This pattern beganto change in the 1960s due to three factors: amajor infrastructure build up (roads,establishment of planned rural settlements,airports, and hydroelectric dams); the concessionof subsidized credit channeled primarily to large-scale ranching; and the establishment of a freeport in the city of Manaus. These initiativesprovided a powerful incentive for immigration,opening up extensive areas for settlement

primarily along the southern margins of the regionand in the major urban centers. Starting in themid-1990s, infrastructure investments by the stateand federal governments have primarily targetedthe paving of existing roads, developing ports, andbuilding a pipeline for hydrocarbons. Marketdemands and these investments have resulted inincreasing human activity in the BrazilianAmazon.

This new wave of investments has also fuelleddisputes over land tenure and forest degradation.As of 2003, about 47 percent of the land in theLegal Brazilian Amazon was public, but withunclear tenure status (Lentini et al. 2003).Conflicts arise because enforcement of land

2BACKGROUND AND TRENDS IN HUMAN

OCCUPATION IN THE BRAZILIAN AMAZON

tenure regulation is weak, and land titling is a long

process. In addition, government land agenciescan expropriate properties considered non-productive, leading to invasion of titled privateproperties or properties undergoing titlingprocesses. In 2002, 26 rural labor leaders wereassassinated in land conflicts in the region (CPT2003). Premature deforestation occur becausepeople can claim ownership by demonstrating thatthey have lived on and worked the land for at leastone year.

These conditions are behind a number oftrends and dynamics that change the landscape ofthe Brazilian Amazon. To understand thesedynamics, it is important to understand the forcesdriving each indicator that was analyzed in thisstudy.

DEFORESTATION

Forest clearingor deforestationhasincreased over time (see deforested areas inFigure 1).Cattle ranching is the most commonland use in deforested areas throughout the region(IBGE 1995). Areas deforested for cattle ranchingoccupied about 70 percent of the total deforested


27/86


26

area by 1995 and pasture area continues to grow.Lower land prices and slightly higher productivitymake mid- and large-scale pastures more lucrativein the Amazon than in other regions of Brazil(Margulis 2003; Arima et al. 2005).2 Moreover,low-density ranching (i.e., less than one head perhectare) offers a lower financial risk than soybean,rice, or corn production (Schneider et al. 2002).

The cattle herd of the Brazilian Amazon grewfrom about 27 million head in 1990 to 64 million

head in 2003, or a mean annual increase of 7 percent.In 2003, 35 licensed meat production plants and 16licensed milk production plants were concentratedprimarily in the southwestern and northeasternportions of the Brazilian Amazon (see Figure 2). In2000, 87 percent of meat produced in the region wasexported to other regions and the rest was consumedregionally (Arima et al. 2005). The location ofslaughter houses and deforestation and distributionof cattle herd (Arima et al. 2005) indicates thatranching is expanding in eastern Par, Mato Grosso,

Tocantins, and Rondnia. Together these statescontained 86 percent of the regions herd in 2003(Arima et al. 2005, based on IBGE data).

Annual crop areas (soybeans, rice, and corn) haveseen notable growth. The area cultivated with annualcrops increased from about 5 million ha in 1990 toabout 8 million ha in 2002 (IBGE 2003b). Growthhas been concentrated in relatively flat and dry zonesalong the eastern and southern margins of the

Brazilian Amazon, usually on already deforestedpastures or in areas originally covered by shrub-savanna vegetation (cerrado) in the states of MatoGrosso and Maranho. In 2002, these two statescontained 83 percent of the area planted to soybeans,rice, and corn in the Legal Amazon (IBGE 2003b).

The prospects for significant expansion ofannual crops in the wetter parts of the BrazilianAmazon are uncertain. As rainfall increases so dothe incidence of pests and diseases, and the highrainfall impairs mechanized harvesting. Forestfallowsindicating areas of low agriculturalpotential and subsequent high costs of chemicalfertilizersreflect this trend. In 1995 forestfallows represented about 8 percent of the totalcultivated area in drier zones of the region, butthey covered 28 percent in wetter zones

(Schneider et al. 2002).

The planting of intensive annual crops(mechanized agriculture) in areas formerly used aspastures is displacing cattle ranching from themargins to the core of the region. An indicator of thisdisplacement is the increase in land values along theroad that connects Cuiab and Santarm (BR-163)in the state of Par: between November 2001 andApril 2002 the land value of pastures and forestlandsincreased, respectively, by 29 percent and 250

percent (estimates based on data from FNP 2002).

Despite legal restrictions, deforestationroutinely impacts environmentally sensitive areas.Brazilian law requires that landowners protectriparian forests and conserve 80 percent of forestcover on their properties as legal reserves. Yetenforcement remains a challenge. A study ineastern Par documented illegal removal of nativeforests in 60 percent of riparian forests along

pastures and other deforested areas (Firestoneand Souza 2002). Government officials in MatoGrosso found 71 percent of 1,600 rural propertiesvisited in 2000 had violated forest lawsincludingdeforestation of riparian forests and clearing oflegal reserves (Souza and Barreto 2001).


28/86

HUMANPRESSUREONTHEBRAZILIANAMAZONFORESTS

27

This map was created by overlaying the INPE map of deforestation as of 2001 over the map of vegetation by IBGE (IBGE Scales of maps: Deforestation (1:250,000). The scale of the vegetation map was 1:2,500,000. The scale of this which aggreforest is 1:1,000,000.

FIGURE 1 VEGETATION AND DEFORESTATION IN THE BRAZILIAN AMAZON AS OF 2001


29/86

HUMANPRESSUREONTHEB

RAZILIANAMAZONFORESTS

28

FIGURE 2 DISTRIBUTION OF TRANSPORTATION NETWORK, TIMBER PROCESSING CAPACITY, ANPROCESSING CAPACITY

Authorized milk and meat production plants are concentrated primarily in the states of Par, Rondnia, and Mato Grosso, ththe largest cattle herd sizes in the Legal Amazon.Sources: Lentini et al. 2003 (Timber processing centers); IBGE 2005 (tranetwork); www.ruralbusiness.com.br/industria.asp?secao=3 (Slaughter and dairy houses).


30/86


29

Data source: IBGE 2003a

FIGURE 3 POPULATION IN THE LEGAL AMAZON

BETWEEN 1950 AND 2000

URBAN ZONES

Between 1960 and 2001, the total humanpopulation of the Legal Amazon increased fromabout 4 million to over 20 million (IBGE 2002).The urban population almost tripled, growingfrom approximately 5 million to 14 millionbetween 1980 and 2000. By contrast, the ruralpopulation began to decline after 1991 (seeFigure 3). The growth in urban areas has led todeteriorating environmental quality and living

conditions. The expansion of highly extensiveranching is linked to urban and nationalmarkets and appears to be a far more powerfuldriver of deforestation than rural populationgrowth. Figure 4 shows the location of the 450municipal seats in the Brazilian Amazon,mapped as of 1997.

AGRARIAN REFORM SETTLEMENTS

Since the late 1970s, landless and urbanpoor have pressured the government for lands.The federal governments Agrarian ReformInstitute (INCRA) grants landless familiesrights to use land holdings as part of agrarianreform projects. The average growth offamilies in the Legal Amazon participating inagrarian reform projects was 52,500 familiesper year between 1994 (161,500 families) and

2002 (528,571 families) (see Figure 5). Eachfamily has user rights to holdings between 50and 100 ha.

The federal government also providessubsidies to agrarian reform settlers in the formof food allowances, money for housing and creditat reduced interest rates. Combined with the


31/86

HUMANPRESSUREONTHEB


30

A 20-km wide radius surrounds the location of the 450 municipal seats in the Brazilian Amazon, to estimate the extent of from these centers. See Section III for details. Source: IBGE 1997.

FIGURE 4 MUNICIPAL SEATS AND URBAN ZONES


32/86


31

Data source: INCRA 2002

FIGURE 5 AREA OF AGRARIAN REFORM SETTLEMENTS ANDNUMBER OF PROJECTS IN THE LEGAL AMAZONBETWEEN 1995 AND 2002

adjudication of legal rights, this makes agrarianreform settlers more prone to deforest than smallscale settlers elsewhere (Wood et al. 2003).

Timber sales also make the initial occupation ofsuch projects attractive for landless people.3

However, after the depletion of timber resources,household income tends to be relatively low.4

Thus, many families abandon or illegally sell their lotsto seek new settlement areas or migrate to urbancenters. An estimated 50-60 percent of land inagrarian reform plots in southern Par has beenillegally sold (Agncia Estado 2004). Some of this landbecomes consolidated in larger land holdings, whichtend to be more economically efficient and profitable.

As of 2002, 8 percent of the total land area inland reform settlement projects in the AmazonBiome had been created to give legal land userights to rubber tappers and Brazil nut collectorsafter Chico Mendes assassination. Most of

these projects (9 of 14)called agro-extractivewere created in the State of Acrewhere Mendes lived. In theory such projects

would be more likely to be used for forestmanagement and small agriculture.Nevertheless, overall, the federal government

acknowledges that land reform has led toenvironmental and social losses, deforestation,and abandonment and subsequent concentrationof land ownership (Presidncia da Repblica2004). Figure 6 shows the location and area ofagrarian reform settlements as of 2002.

FOREST FIRES (FIRE ZONES)

Although many Amazonian forests have astrong capacity to resist burning, uncontrolledfires are a growing problem in the BrazilianAmazon (Nepstad et al. 2004; Cochrane 1999).


33/86

HUMANPRESSUREONTHEB


32

Source: INCRA 2002. Scale of Agrarian Reform Settlement: 1:100,000.

FIGURE 6 AGRARIAN REFORM SETTLEMENTS ESTABLISHED AS OF 2002


34/86


33

Burned forests are greatly susceptible to recurrentfires, which in turn can be more severe inintensity and impact (Cochrane 1999). Fire is theprincipal tool used to clear land for planting soonafter deforestation, and thereafter to maintainpasture. Fire accidentally escapes fromagricultural lands to forested areas, mostly tologged areas that become more susceptible toburning. Half of the forest fires in the Amazon areaccidental (Nepstad et al. 1999) and can occureither where human occupation is consolidated or

in occupation frontiers. Data is available on thedaily incidence of fire. Between 2000 and 2002 thenumber of fires nearly doubled from 22,000 peryear to almost 43,000 per year, indicating theacceleration of human occupation (see Figure 7).

MINING

Mining activities often do not involve directclearing of large forest areas, but they can serve as

catalysts of deforestation and timber harvestingbecause they are associated with road building,capital accumulation, and immigration (Bezerra etal. 1996). For this reason, maps on miningoperations can be used to flag areas of incipientand potential human activities (see Figure 8).

According to Brazilian law, mining takespriority over any other surface or sub-surface landuse. Should areas allocated for mining activities ormineral exploration become economically viable,

improved infrastructure, transportation, and otherservices could spark rapid in-migration and forestclearing.

In addition to legal mining operations,informal miners (wildcat mining) are also a

source of human pressure in the BrazilianAmazon. In the early 1990s there were about 1million gold miners (garimpeiros) in morethan two thousand mining camps (Pinto1993). Gold mining exerts directenvironmental impacts from forest clearingfor the mine and the adjacent mining camp(Bezerra et al. 1996), and indirect impactssuch as soil erosion and mercury pollution(Mathis and Rehaag 1993). Construction ofinfrastructure and accumulation of capital

provide a basis for other land-use activitiesand additional frontier expansion. Thisphenomenon occurred in the Tapajs Riverbasin of western Par, where around 245 goldmining camps used to employ roughly 30,000people in the early 1990s (Bezerra et al.1996). The gross value of the gold extractedreached US$110 million per year, whichhelped to finance land conversion to non-forest uses such as cattle ranches. When goldresources have been depleted, unsuccessful

miners frequently seek agrarian reformsettlements or move to urban centers.

LOGGING

Logging has been a major catalyst forsettlement in the Brazilian Amazon becauseloggers open roads and use navigablewatercourses to reach native forests. In 2004

approximately 3,100 wood-processing mills(sawmills, veneer and plywood mills) processed24.5 million cubic meters of logs; 36 percent ofthis timber was exported to other countries andthe remainder consumed in Brazil (Lentini etal. 2005).


35/86

HUMANPRESSUREONTHEB


34

Fire Zonesforest fire locations surrounded by a 10 km-wide bufferare used to estimate the zone of incipient human aSection III for details. Here, areas with exclusively recent fires zones show new frontiers of occupation. Source: INPE 200

FIGURE 7 FIRE ZONES, 1996-2002


36/86


35

FIGURE 8 AREA ALLOCATED FOR MINING OR MINING EXPLORATION AS OF 1998

Source: DNPM 1998.


37/86


36

The environmental and ecological impacts oflogging have only been partially evaluated, andthey can be extremely variable due to the variety oflogging methods and the occurrence of secondaryimpacts. Nevertheless, loggers have openedthousands of kilometers of roads in public andprivate areas that have become key channels forfurther colonization (Verssimo et al. 1995;Greenpeace 2001; Brando and Souza 2006).

Selective logging is widespread in large areas

and can cause light to severe damage. Unplannedlogging generates greater amounts of slash andopens larger gaps in the forest canopy thanplanned harvesting operations, making forestsmore susceptible to fires that originate in areasused for shifting cultivation or pastures(Verssimo et al. 1992; Verssimo et al. 1995; Uhlet al. 1991; Holdsworth and Uhl 1997; Johns et al.1996; Gerwing 2002). Impacts from licensedlogging operations are also negative when licensedloggers neglect to adopt the approved

management procedures.5 Some timbercompanies and communities have adopted bestforest management practices, which in theBrazilian Amazon is mostly equated with theadoption of green certification, since thecredibility of non-certified operations is low.However, the total area of green certified timberproduction in the region remains small. As ofNovember 2005, there were 12,619 km2 of FSCcertified lands (FSC 2005). This land is

equivalent to only 3.4 percent of the estimatedtotal area needed to supply the annual timberharvest in the Brazilian Amazon.

The total area affected by various timberharvesting methods in the Brazilian Amazon is

unknown and datasets are temporally and spatiallyincomplete. The current status of licensed loggingoperations, for instance, is unclear, since the laststatus report released by IBAMA was based on datafrom 2001. Rough estimates of area impactedannually ranges from 10,000 to 20,000 squarekilometers, with uncertainties from 17 percent to100 percent (Nepstad et al. 1999; Matricardi et al.2001; Cochrane 2000). Of special interest is themost recent analysis (Asner et al. 2005) that usessatellite imagery between 2000 and 2002 to identify

areas affected by selective logging.

Comparing the information that is available,such as the location of logging permits issued bythe government (see Figure 9) and maps of loggedforests, could be used as a rough surrogate toidentify areas where timber harvesting isoccurring and how harvesting is distributed.Systematic interpretation of satellite imagery,combined with fieldwork, is a promising tool toevaluate the extent and impact of logging at a

regional scale (see Figure 10).

ROADS

A number of studies demonstrate that roads, eventhose opened temporarily by loggers and gold miners,facilitate subsequent settlement. A few examples:

Timber companies in search of mahogany

(Swietenia macrophyla) were the majorbuilders of logging roads in southern Parduring the 1980s; by 1992 this networkextended nearly 3,000 km (Verssimo et al.1995) and has continued to proliferate since(Greenpeace 2001).


38/86


37

FIGURE 9 LOCATION OF LOGGING PERMITS IN 2000

Source: Greenpeace 2000. Logging permits for forest management plans. Timber harvest from clear cut (deforestation) i


39/86

HUMANPRESSUREONTHEB


38

A pilot study in eastern Par (Souza et al. 2003) found that approximately two-thirds of the forest in a 3,700-km 2 area had severely degraded by logging and fire. This study demonstrates the potential of remote sensing and fieldwork to systematiimpacts of logging throughout the Brazilian Amazon.

FIGURE 10 LAND COVER CLASSIFICATION IN EASTERN AMAZONIA


40/86


39

During the 1980s, gold miners and miningcompanies also opened informal roads insouthern Par (Mertens et al. 2002). Settlersadvanced along logging and mining roadswherever agriculture or ranching becamefeasible (Verssimo et al. 1995; Mertens et al. 2002).

Timber companies, settlers, and localgovernments have also constructed feederroads along the Transamazon Highway in Par,northeastern Mato Grosso, and southernAmazonas (Rodgers 2003; Greenpeace 2002),

which could explain the high concentration offorest fires in these areas.

Of special concern are unofficial roadsroadsbeing built without the planning and authorizationrequired by law (see Box 4). In some cases,

unofficial roads serve only for one-shot extractionof high-value resources, such as mahogany andgold, in areas isolated from official infrastructure.Loggers, ranchers and miners have opened a vastand growing array of unofficial roads, enablingtemporary or permanent human occupation overextensive areas of the region. The vastproliferation of unofficial roads suggests an evenwider impact and more rapid change in theBrazilian Amazon.

The intensification of human pressure dependslargely on the continuation of investment intransportation infrastructure. This is happening insome areas because early settlers who buildinformal roads usually pressure government tomaintain and improve such infrastructure.


41/86


40

Thousands of kilometers of roads are beingbuilt in the Brazilian Amazon without theplanning and authorization required by law.Unofficial roads include private roads and illegalor informal roads. Until recently, the growth andextent of unofficial roads in the BrazilianAmazon had not been documented. With support

from WRI and others, Imazon identified,mapped, and quantified the length and growth ofunofficial roads in the Central-West region of theState of Par (the Midlands, or Terra do Meio),an area that represents about half of the State.Analysis relied on visual interpretation ofsatellite imagery. Almost half of the Midlands areunclaimed public lands, while a network ofprotected areas extends over the other half(IBGE 1997; ISA 1999).

Three types of roads from Landsat satelliteimages were identified for three periods (1985-1990, 1991-1995, and 1996-2001):

Visible roads: continuous linear featuresvisible to the human eye in the images;

Fragmented roads: non-continuous linearfeatures visible to the human eye. Fragments

may be traced to connect the hiddenstretches to those that are visible in theimage.

Partially visible roads: linear features thatare not explicit in the images and may onlybe identified based on their context andspatial arrangement (i.e. adjacent deforested

areas).

Imazon found that 80 percent of the roads(~ 21,000 km) in the Midlands are unofficial.Moreover, almost 60 percent of these corridors arein unclaimed public lands penetrating largeblocks of forests that are potentially appropriatefor the creation of protected areas. The resultssuggest that protected areas slow the advance ofunofficial roads because average growth rates forunofficial roads inside protected areas were three

times lower than those outside protected areas.

Building on this experience, Imazon isextending the mapping of unofficial roads intoother states of the Brazilian Amazon. A moredetailed description of the mapping methods, theresults and implications can be found in Brandoand Souza (2006).

BOX 4. MAPPING UNOFFICIAL ROADS IN THE MIDLANDS OF PAR


42/86


43/86


42

other human activities. Fire zones containsmall areas (mostly less than 3 ha) that weredeforested and burned, but were notincluded in the INPE deforestation mapbecause they are smaller than the mappingthreshold. Fire zones can also flag loggedforests into which wildfires escaped fromnearby deforested areas. In summary, thiszone is a crude indicator of areas underincipient pressures from hunting, smalldeforestation, forest fires, and logging.

The validity of using the 10-km thresholdwas tested when examining the relationshipbetween human pressure, roads, and logging(see below). The results seem to support thethreshold: half of the authorized loggingoperations were within fire zones;and,inareas where detailed information on roadswas available, 76 percent of the fire zone areawas accessible by roads or navigable rivers.(See discussion of logging in Section IV).

Although forest fire data were available onlyafter 1996 and the database is incomplete for1996 and 1997-98, fire zones are likely tocapture a significant portion of the incipienthuman pressure.6 In areas with older signs ofincipient occupationfor example, whereforest fires occurred prior to 1996deforestation either increased and, hence, theareas were included in the INPE deforestation

map; or occupation remained incipientthatis, small farmers continue to practice smallscale slash-and-burn agriculture. In this case,new forest fires were captured in the vicinity ofold cultivated areas and were included in the10-km buffer around fire zones.

Areas with exclusively newer forest fires mayindicate zones where the occupation frontieris expanding. To show such zones, some ofthe maps distinguish exclusively newer firezones (2000-2002) from older fire zones(1996-1999). To separate such zones, firezones with both old and new forest fires wereclassified as old fire zones.

Areas allocated for mineral research andmineral reserves as of 1998 (Capobianco etal. 2001). The areas licensed for mineral

prospecting are not necessarily active, butthey may become so if minerals of interestare found (see Figure 8).

This sequence indicates progressively lowerhierarchy of data layers. Thus, maps with the mostprecise information and showing the greatestdegree of transformation in the natural vegetationtook precedence over information that was lessprecise or that indicated less pressure. Adeforested area overlapping areas of influence

from urban zones or an area licensed for mineralresearch, for instance, was classified asdeforested.

In the absence of a map of all logged forests, amap of the location of official permits for loggingin 2000 was overlaid on the human pressure mapto provide a crude indication of regions pressuredby timber harvesting (see Figure 9). Thegeographic coordinates of logging permits were

compiled by Greenpeace Brazil using data fromIBAMA.

In addition, a preliminary comparison betweenthe map of human pressure and the map depictingselective logging between 2000 and 2002 (Asner et


44/86


43

al. 2005) was performed. The Asner study is themost recent analysis of selective logging throughoutmost of the region using satellite imagery.

HUMAN PRESSURE IN PROTECTED AREAS

The map of human pressure was overlaid onexisting datasets of protected areas:

Existing protected areas as of 2004 (ISA

2004; Viana and Valle 2003). Priority areas for conservation outside

protected areas. These areas wereidentified by overlaying the map of priorityareas for conservation (Capobianco et al.2001) on the map of protected areas.Although the map of priority areas forconservation is the most updated compilationof available inventories of taxonomic groupsin the region, the map is incomplete due tolimited or non-existing sampling in more

remote zones. As a result, the assessment ofhuman pressure may underestimate thepriority areas for conservation.

Areas identified with potential for theestablishment of public productionforests (Verissimo et al. 2000).

The hypothesis that the human pressure onprotected areas is influenced by the proximity ofthese protected areas to roads was tested by

estimating the Spearmans non-parametriccorrelation coefficients between the proportions ofprotected areas within distance intervals fromroads (< 25 km, 25-50 km, and >50 km) and theproportions that were (1) deforested and (2)within a 10-km radius around fire zones.

HOW THE RELATIONSHIP BETWEENROADS AND HUMAN PRESSURE WASANALYZED

Three approaches were used to assess theinfluence and relationship between roads andhuman pressure:

Measuring the cumulative distribution(percent) of forest cover and human pressurewithin intervals of 10 km from each side of

the official roads existing by 1999 (IBGE2003a).

Analyzing the association between officialand unofficial roads in northern Mato Grosso(76,000 km2) and central and southwesternPar (546,000 km2). This was done bymeasuring the cumulative distribution(percent) of unofficial roads within 10-kmintervals from each side of the official roadnetwork.

Measuring the area under incipient human

occupation that was intercepted by at leastone transportation route (official or unofficialroads or navigable rivers) in areas where thecomplete road network was available (Souzaet al. 2004).

The analysis of human pressure is part of aneffort coordinated by Global Forest Watch (GFW)to examine the extent and impact of humanactivities in important forested regions of the

world. Analyses elsewhere include: Mapping forest intactness by identifying

large forested landscapes least impacted byhuman activities visible in satellite imagery inboreal forests of Canada, Alaska, Russia,


45/86


44

Fennoscandia and Alaska, and in costaltemperate rainforests of Canada, Alaska andChile (Aksenovet al. 2002; Leeet al. 2003;Hjek 2000; Strittholtet al. in preparation,Verscheureet al. 2002).

Mapping forests accessed throughtransportation routes in tropical forests ofIndonesia (FWI/GFW 2002) and CentralAfrica (Minnemeyer 2002); tracking the

expansion of logging roads to monitor forestdevelopment and to identify potential illegallogging (Van Polet al. 2005).

Mapping methods used in these other global,regional and national analyses attempt to establishan overall consistency at the same time that theyare sensitive to local realities (e.g. data availabilityand ecological differences).


46/86


45

As of 2002, approximately 47 percent of theBrazilian Amazon shows some indication of

human pressure (see Table 1 and Figure 11).Areas under pressure from human settlementsrepresent almost 19 percent of the study area,while areas subjected to incipient human pressurerepresent 28 percent (see Figure 12). Of the areasunder pressure from human settlements,deforested areas accounted for 11 percent of thearea, urban areas for nearly 6 percent, andagrarian reform settlements for nearly 3 percent.Fire zones accounted for the majority of the areassubjected to incipient human pressure.

The relative area of non-forest vegetation undersome type of pressure (66 percent) is much higher

4HUMAN PRESSURE IN THE

BRAZILIAN AMAZONRESULTS AND DISCUSSION

than the percentage of forest vegetation underpressure (44 percent). About 14 percent of all land-

cover types are non-forest vegetation in the BrazilianAmazon, according to the vegetation map from IBGE.

Examining the distribution of human pressurebetween forested areas and non-forested areas, mostareas of non-forest vegetation under some type ofpressure are located in the eastern and southernportions of the Amazon biome and in eastern Roraima(in the upper northern Amazon). The pressure onnon-forest vegetation is relatively smaller in

campinaranas (or heath forests)7 of the northernAmazon and savannas of southern Roraima. In theseareas pressure is mainly indicated by scattered firezones mainly alongside navigable rivers.

TABLE 1 HUMAN PRESSURE IN THE BRAZILIAN AMAZON

Areas under Pressure from Human SettlementsDeforested AreasUrban ZonesAreas of Agrarian Reform Settlements

Areas Subjected to Incipient Human PressureFire ZonesMining

TOTAL**totals may not add up due to rounding

Brazilian Amazon

Percentage ofTotal Area

Percentage ofForest Vegetation

Percentage of Non-Forest Vegetation

Type of Pressure

191163

2727


47/86

HUMANPRESSUREONTHEB


46

FIGURE 11 HUMAN PRESSURE IN THE BRAZILIAN AMAZON - ALL INDICATORS


48/86


47

FIGURE 12 TWO TYPES OF HUMAN PRESSURE

Areas subjected to incipient human pressure include fire zones and areas allocated to mining. Areas under pressure fromsettlements include deforested areas, urban zones, and lands allocated for Agrarian Reform Settlements.


49/86


48

AREAS UNDER PRESSURE FROM HUMANSETTLEMENT

Areas under pressure from human settlementsare heavily concentrated near roads; they aremostly located in the so-called arc ofdeforestation, stretching from eastern andsouthern Par through northern Tocantins andMato Grosso, cutting through Rondnia, andreaching into eastern Acre. Other major areas ofpressure from human settlements occur along the

TransAmazon Highway in Par, the Amazon Riverbetween Manaus and Belm, the Cuiab-Santarm highway near the city of Santarm, andaround the main urban centers in Roraima andAmap (Figures 11 and 12).

Deforestation

The results show that about 11 percent of theBrazilian Amazon was deforested as of 2001.Deforestation in the Brazilian Amazon is different

than the estimated deforestation for the LegalAmazon. The Legal Amazon contains greatlymodified areas in its southern and easternboundaries and these southern and eastern areasare not considered part of the Brazilian Amazon;deforestation in the Legal Amazon is estimated at15 percent for 2001. (See Box 3 for an explanationof the overlap between the Brazilian Amazon andthe Legal Amazon.)

Deforestation fragments the landscape andcreates more edges between forests and non-forested areas (Laurence et al. 2000). By 1988, theforest area at risk of edge effect (< 1 km from theforest edge) in the Amazon was about 150 percentlarger than the total area deforested (Skole and

Tucker 1993). Forest edges are affected by solarradiation, wind, and agricultural fires (Cochraneand Laurance 2002). Forest inventory studies haveshown that the biomass of forest edges decreasesdrastically within 100 meters of the edge(Laurence et al. 1997). The species diversity andcomposition also change in the forest edges andthe edge effect could contribute significantly tothe emission of green house gases such as CO2

(Laurence et al. 1997). All these factors lead to amore degraded forest environment within forest

fragments and forest edges.

Most of the forest fragments in the BrazilianAmazon are found within settled areas. In areas ofincipient human pressure, forest fragments areless common, but the length of forest edge tendsto increase. Additionally, as more unofficial roadsare built in the incipient human pressure frontier,large forest fragments tend to become isolated. Inthis report, we have not quantified the overalleffects of forest degradation due to forest

fragmentation.

Urban Zones and AgrarianReform Settlements

Approximately 6 percent of the BrazilianAmazon lies within urban zones; this is, thelocation of municipal seats and the 20-km bufferaround them. The total area of agrarian reformsettlements accounts for nearly 5 percent of the

Brazilian Amazon and about half of the agrarianreform settlements are either forested or outsideurban zones (see Figure 11). In 2001,deforestation in the agro-extractives agrarianreform settlements (PAE, for the acronym inPortuguese) was much smaller (2 percent) than in


50/86


49

conventional land reform settlements (43percent). The apparent success of PAEs to haltdeforestation, however, may be misleading since,until recently, most of them were in remote areas.In the State of Acre, for instance, INCRA officialshave reported that with improvements ininfrastructure (mostly road paving) deforestationand illegal sale of land also occurs in PAEsettlements (Pgina 20 2005). A highly publicizedexception is the Chico Mendes PAE, where settlersare managing the forest for timber production. In

2002, 1,900 ha of forest management in the ChicoMendes PAE gained FSC certification (FSC 2005).This success has been possible due to extensivesupport from Acres state government and NGOs,and because the area allocated per family isrelatively large. It remains a challenge, however, toreplicate this success. It is relatively difficult togarner such a high level of support for all the landsettlement areas and most of the per-family landallocations in agrarian reform settlements (usuallyless than 100 hectares) are not large enough to

generate sufficient income from sustainableforestry. (See analysis in Menezes 2004).

AREAS UNDER INCIPIENTHUMAN PRESSURE

In contrast to areas under pressure fromhuman settlements, a greater proportion ofincipient human pressure was beyond 30 km from

official roads (See Figure 11). Fifty-four percent ofthe area of old fire zones, two thirds of recent firezones, and 96 percent of mining reserves/licenseswere beyond this threshold. Human presence insuch areas is possible due to navigable rivers and agrowing network of unofficial roads.

Fire Zones

Approximately 28 percent of the BrazilianAmazon was subjected to incipient humanpressure associated with fire zones, measured as a10-km radius around a forest fire. This arearepresented over half of the total area in whichsome form of pressure was detected.

The presence of fires and roads within firezones suggests the presence of several incipient

human activities such as deforestation of smallplots (mainly less than 3 ha), logging, extraction ofnon-timber forests products, and hunting.

Different patterns of fire zone distributionseem to indicate different densities of occupation.Nearly two thirds of the area of the fire zones isaggregated (i.e., more than three overlapping firezones). These aggregated areas extend beyond themargins of major settlements in Rondnia, MatoGrosso, and Par. Furthermore, areas subjected to

incipient human pressure around fire zones alsooccur in northern Roraima, eastern Amap, andsoutheastern Acre and Amazonas. Thisconcentration suggests that new pressure isoccurring mostly in proximity to areas with denserinfrastructure and population.

About 34 percent of the fire zones occur eitherin low densities or isolated groups (three or lessoverlapping fire zones). These low-density fire

zones indicate at least two distinctive types ofpressure. First, those located closer to settledareas (e.g., deforested and urban zones) may markthe beginning of settlement in forested areasalong or in the vicinity of newly opened roads (seeexample in Verssimo et al. 1995). Deforestation


51/86


50

far from markets, where farming profitabilitytends to be small, is partly explained by landspeculation. Colonizers establish pasture inthese zones expecting to gain from futurevaluation of the land when infrastructuredevelops. Land speculation is also stimulated bythe fact that colonizers can claim public land astheir property by showing evidence that they arecultivating the area.

Many of the smaller clusters of fire zones

occur in isolated areas along rivers in the statesof Amazonas, Acre, and Par. Traditionalpopulations in these areas practice shiftingsubsistence agriculture in small plots (typicallyless than 2 hectares) that are cleared, burned,and used for planting cassava, corn, and othercrops, after which they are left fallow. Localresidents also hunt and harvest wood and non-timber forest products such as rubber, Brazilnuts, seeds, and fruits from the surroundingforests. The impacts of this land-use pattern

vary greatly but are far less than those generatedby intensive logging and large-scaledeforestation, although it can lead to localizedextinctions of game species (Nepstad, et al.1992; Redford 1992).

Mineral Reserves and Areas Licensed forMineral Prospecting

Less than 1 percent of the Brazilian Amazonwas exclusively under incipient human pressurefrom mineral reserves and areas licensed formining. The total area legally allocated for miningis equivalent to approximately 2 percent of theBrazilian Amazon.

Logging

The results of our analysis suggest that mostlogging has occurred within the areas under pressurefrom human settlements and areas subjected toincipient human pressure. Most of the authorizedlogging operations as of 2000 overlap with the map ofhuman pressure, and only 15 percent of theoperations are in areas free of other indicators (Table2). The distribution of logging authorizationscorresponds to the distribution of wood production:approximately 80 percent of the permits are in Par,

Mato Grosso and Rondnia states which harvested 93percent of the total timber harvested in the BrazilianAmazon in 2001 (Lentini et al. 2003). There was nooverlap between logging operations and mineralreserves and areas licensed for mineral exploration.


52/86


51

Non-Forest VegetationForests under No Human PressureDeforestationUrban Zones

Agrarian Reform SettlementsFire ZonesTotal

TABLE 2 OVERLAP BETWEEN LOCATION OF LOGGING OPERATIONS AND OTHERINDICATORS OF HUMAN PRESSURE

Human Pressure Indicator or Land Cover Type

Number of LoggingOperations Found in Each

Indicator Category orLand Cover Type

Percent (Rounded) ofLogging Operations Foundin Each Indicator Category

or Land Cover Type

285

11157

32293580


53/86


52

FIGURE 13 CUMULATIVE AREA (PERCENT) OF HUMAN PRESSUREIN RELATION TO DISTANCE FROM OFFICIAL ROADS

As is the case for deforestation, the distribution

of unofficial roads is strongly influenced by thepresence of official roads. In south-central Parand northern Mato Grosso, for example, 82percent of the total length of unofficial roads iswithin 50 km of official roads. Considering thesetwo regions, unofficial roads appear to beassociated primarily with the consolidation ofsettlements: 62 percent of the length of unofficialroads crosses areas under settlement; anadditional 27 percent of the length of unofficialroads is located within fire zones; and 11 percentcrosses forested areas without other signs ofhuman pressure.

Our analysis confirms the effectiveness of usingfire zones as a surrogate to identify incipient

human pressure in areas where transportation data

is incomplete. In the areas where detailed maps ofthe road network are available, the majority of thefire zones are accessible by unofficial roads andnavigable rivers. In the areas of south-central Parand northern Mato Grosso, 76 percent of the firezones are accessible, 45 percent of them exclusivelyby unofficial roads; 17 percent by unofficial roadsand other types of access; 13 percent by navigablerivers; and 1 percent by a combination of otherforms of access (see Figure 14). An estimated 24percent of the area of fire zones shows no evidenceof access. In part, this could be due to the fact thatunofficial roads have not been mapped for theentire Brazilian Amazon. Figure 15 shows theoverlap between unofficial roads human pressure insouth-central Par.


54/86


53

FIGURE 14 TYPES OF ACCESS TO FIRE ZONES IN SOUTH-CENTRALPAR AND NORTHERN MATO GROSSO

FIGURE 15 UNOFFICIAL ROADS, HUMAN PRESSURE, AND INDIGENOUS LANDS

Ilustrates the overlap between official and unofficial roads, areas under human pressure,and Indigenous Lands.


55/86


56/86


55

5HUMAN PRESSURE AND PROTECTED AREAS

RESULTS AND DISCUSSION

This section presents the extent of human

pressure in current protected areas and in priorityareas for the creation of new protected areas. Thisincludes areas both for conservation and forpublic production forests. This section ends witha summary of risks and opportunities for thecreation of new protected areas.

HUMAN PRESSURE IN EXISTINGPROTECTED AREAS

Overall,human pressure in protected areas ismuch smaller than in un-protected areas (seeFigure 16). The proportion of areas underpressure from human settlements in non-protected areas (28 percent) was almost seventimes higher than in protected areas(approximately 4 percent).

The area deforested in protected areas was

equivalent to about 1.3 percentmuch less thanthe 17 percent outside protected areas. Ot

Human Pressure Amazon

Documents

Transcript of Human Pressure Amazon